Considerations in Engine Cowl Design

[RFguy]

There are different ranges of those small radiators. No, I won't be using those radiators specifically, they dont have sufficient pressure capability. just an example of what is out there .

 

But there are a range of like things for cooling electronics in harsh vibration environments.

but The motorcycle motoX radiators are the most likely - although it really requires a custom job because inevitably the inlets and exits are in the wrong place, and we want crossflow and most of the the motoX  are top to bottom, (some are X flow) . Still, brazing or tigging ally pipes into the aluminium is not difficult.

 

I dont have any practical feel for how the shapes of the tanks on a crossflow influence distribution of the coolant. Does anyone know ?

You see tapers on the tanks at the sides in some of the higher performance crossflows.

 

The radiator supplied with the LCH kit is a crossflow from a CBR1300 I think.

 

One can go deep core radiators, stacked but any of that is lower in efficiency than just more surface area as of course the air heats up as it passes the first cores. But better than nothign once you are out of available surface area and cant forget there is plenty of static pressure to push through multi cores. This is not thermo fan territory.

Temperature difference is king.

 

ANyway Skip, the 4 and 5" dia  lines up to a meter long are suitable for piping airflow with little pressure loss in our application and speed. Cant go below 4" ID. You could pipe the air directly to the radiator if you have trouble. I reckon getting that seal on core cowling will take a bit of work.

The seal has to be good (stable and rigid ) under a couple of inches of water pressure, which really is not many psi fractions.

But, let's see how your bird goes.  I think I calculated 18 liters of coolant was enough to start up, do a circuit and taxi and shutdown without exceeding 100 deg C. (in case of having no coolant cooling- just energy to heat up the coolant).

 

 

Edited January 19, 2022 by RFguy

[skippydiesel]

52 minutes ago, RFguy said:

--------------------------we want crossflow and most of the the motoX  are top to bottom, (some are X flow) . --------------------------

 

I

Why??

 

Please explained/expand.

[RFguy]

In my case I have generally, lots of space  in the horizontal plane, and little in the vertical plan, so cross flow will give me the most useful geometry from a number of cores perspective .

 

From what I read, getting even flow distribution  is more difficult the more pipes you have/longer the tanks. in the wide aspect cross flow, the tanks are short in length (height), and the number of pipes is less. 

I'm sure there is a expert here that can pipe up on this. 

 

If the resistance of the flow in each tube is low compared to the input pressure than the system should self balance.
IE long tubes in a cross flow will have higher individual pressure drop than short tubes. and thus they'll distribute more evenly.

 

But its not quite so simple- since there is variation in density, volume as the fluid goes from  hot to cold ! more reading required.

Edited January 19, 2022 by RFguy

[RFguy]

I found a good paper to read

"Effect of non-uniform airflow on the performance of a parallel-flow heat exchanger considering internal fluid distribution"
document source : https://doi.org/10.1016/j.applthermaleng.2020.115685

 

Now, interesting  for this crossflow radiator (aspect = 1.38) square uniform side tanks, no special shapes) as the mass flow rate goes up, so does the variation in distribution of flow. and that the tubes in the middle suffer . 

A number of different aspect ratios was tested (width/height ratio), and non uniformity of flow distribution increases with reducing aspect ratio- so square is crap. and because of the pressure drops in the tubes becomign greater compared ot the in/out tank presssures as I'd expected (very similar to electricity and laws of current sharing) .

 

and when you take into account the effects of  non uniform airflow , then the radiator performance will decrease more quickly (and we are talking 4x difference for aspect = 0.7 to aspect = 1.4 ) This is significant. so use wide aspect ratios. 2:1 or better.

 

 

 

 

 

 

Edited January 19, 2022 by RFguy

[skippydiesel]

Ah! BUT orientation of the tank/tubes (ref your statement & my Q) should have very little difference on performance. You can pretty much put a Rotax radiator in any position and get the same efficiency SUBJECT TO GOOD AIR FLOW.

 

This holds good for all radiators - I have converted a Ford Falcon cross flow radiator (tank on each side horizontal tubes) to vertical (tank top/bottom tubes vertical - did the same job as when in in its previous orientation.

 

Radiator orientation has more to do with space availably than physics.

[facthunter]

Geez Fella's, It's not a moon rocket or a race car. There are plenty of 912's around running too cold that need foam over a part of the radiator in winter.. Nev

[onetrack]

There's some useful cooling information (car-based) in the link below. A couple of the major factors in efficient cooling appears to be in fin design and tube shape. 

 

https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.824.1312&rep=rep1&type=pdf

[skippydiesel]

6 minutes ago, onetrack said:

There's some useful cooling information (car-based) in the link below. A couple of the major factors in efficient cooling appears to be in fin design and tube shape. 

 

https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.824.1312&rep=rep1&type=pdf

Makes sense to me.

[onetrack]

You can acquire double or triple pass crossflow radiators, where the tanks have restrictions that divert the coolant flow to keep criss-crossing the core. These double or triple pass radiators are more efficient because they keep the coolant in the airflow for longer.

In too many radiators, the coolant flow is too fast to permit maximum coolant heat transfer. That's why removing a thermostat usually doesn't cure overheating problems, it usually only worsens them.

[skippydiesel]

15 minutes ago, onetrack said:

You can acquire double or triple pass crossflow radiators, where the tanks have restrictions that divert the coolant flow to keep criss-crossing the core. These double or triple pass radiators are more efficient because they keep the coolant in the airflow for longer.

In too many radiators, the coolant flow is too fast to permit maximum coolant heat transfer. That's why removing a thermostat usually doesn't cure overheating problems, it usually only worsens them.

Seem to me that most overheating problems, relate to lack of air flow, rather than coolant "speed".

 

Experimental set ups aside - got a cooling problem -  First check air flow restrictions then  external "core" blockage which  could be anything from damaged fins, insect or dirt build up.

 

Air flow restrictions - body damage/ accessories that partially block air flow - fix/remove. Paper/plastic/straw/etc - manually remove

 

Damaged/flattened fins  can sometimes be partially rehabilitated using a thin flat blade screw driver or even a bit flat metal out of a windscreen wiper GENTLY& PROGRESSIVLY  return the flatted fins to something like their original position.

 

Drive through a swarm of bees/grass hoppers/ blowing chaff/etc and your radiator leading surface is likely to become blocked in whole or part - best to reverse blow with compressed air (access can be a problem requiring removal of radiator).

 

Dust/chaff/insects can slowly build up over time progressively reducing radiator efficacy use above treatment. Sometimes using a brush with air to disturbed the material works best. Worst cases may require "rodding" with a flat wire such as found in some windscreen wiper refills.

 

Be careful about using high pressure water instead of air- very easy to damage radiator fins and sometimes the water will cause the material to swell preventing wash out.

 

After that consider a radiator flush to remove internal coolant flow restriction - the habitual use of tap /dam or contaminated water will eventually block internal cores.

 

Personally - when I have a fully functional radiator, on a ground vehicle/engine, I cover the inlet side with fiberglass fly screen material. The minimal air flow restriction that the fly screen imposes is nothing compared with the above scenarios - Leave the bottom of the fly screen loose, the continually movement in the air flow will tend to "shed" insect bodies & the like. I have not had an overheating problem, due to a contaminated radiator in 45 + years of driving & maintaining.

[RFguy]

So, skippy, I fu***ed up on my calcs a little.

I assumed full power and I assumed high density altitude  like 10k ft.

BUT. you dont get full power at 10k feet !. doh. idiot me. It was in another discussion today and it just twigged by brain. 

 

so the problem gets easier (unless you have a turbocharger) . like my calcs go from 440cm2 > 300cm2 area.  IE air density 1.2kg/cuM for the exercise instead of 0.85kg/cu3.

 

The 450cm2 radiator that rotax these days says to have , that's for engines up to the 914, and of course the 914 has a turbo, and that's probably why, at least roughly, my numbers matched up and led one astray.

Edited January 27, 2022 by RFguy

[RFguy]

Hi SKippy

As I continue to learn about fluid dynamics, some things are now apparent by the numbers...

firstly - 

For liquid cooled aircraft,  using the surface of the leading edges of the aircraft, like using tube forming the wing leading edges,  or the skins being heat exchangers would be fantastic for liquid cooling, and wouldnt require any front area dedicated for cooling... seems a no brainer, but I have not seen anyone doing it.  

 

The leading edge would seem a no brainer, and some aircraft I think the RANS the leading edge is already an alumuminium extrusion. 

The skins, while that only receive boundary airflow (lower velocity etc) , large areas available would make them work !

If I replaced my Jabiru top or bottom cowling with a skin of aluminium with water in behind it, like a flow structure etc (there are many flat, formable aluminium extrusions designed  for intercoolers etc )  , there would be more than enough surface to cool anything.... I dont have time to weld up the tanks on each side of the flat tubes so I'll use draggy radiators.. 

 

Now the bad news :

 

Your small inlet  (suggest min 288cm2) , large  plenum (the cowling) is akin to a hydraulic  amplification.  small  inlet into a large volume. 

This increases pressure of course.

Your exit air mounted radiator will work grandly IF you can get the whole thing sealed, and not otherwise- the high pressure will go hand in hand with sealing challenges.

You'll have some gaps , so the minimum inlet size might need to be increased quite a bit.  In the photos , it look slike a substantial gap will be required around the exhaust pipe. 

 

the next problem you might face is the production of a low pressure region where you want it. This wont be flush with the skin, it will be up a recess/duct. ideally it will work as some thrust.

 

Next thing, parasitic drag is proportional to velocity CUBED . Put a 500cm2 flat plate in the wind at it will consume 10HP at 120 ks !!!

Unfortunately, you are going to have to get air from somewhere to run your radiator, which means having SOME frontal area.

 

 

glen

 

 

[Geoff_H]

I was not very good at fluid mechanics, but one principle has stuck with me, the flow velocity of a fluid over a surface is zero at that surface.  With respect the problem at putting a heat exchanger in the wing leading edge is the velocity at the leading edge will be zero hence:

"Turbulent flow, due to the agitation factor, develops no insulating blanket and heat is transferred very rapidly. Turbulent flow occurs when the velocity in a given water channel is high. ... Laminar flow develops an insulating blanket around the channel wall and restricts heat transfer."

 

Best regards Geoff

[Geoff_H]

Please take not that the cooling air will be blanketed from the wing flow.  This technique is used in high performance gas turbines where a layer of air is run over the first row of blades to blanket them from hot combustion gas. Is

[RFguy]

Hi Geoff

 

I should have written
power required to overcome parasitic drag is proportional to velocity CUBED

instead of 

parasitic drag is proportional to velocity CUBED

 

thanks for the input.  yes that's how they avoid melting N2 blades I gather.

OK, not so efficient idea. but lots of surface area. again considering just boundary layer flows on the wing . 

the surfaces, like the cowling , although having only boundary layer flow, there is oh so much surface area to work with . Your thoughts ?

 

SKippy, have you had time to make any progress ?

 

 

 

 

 

[Bruce Tuncks]

I like the idea of using the wing leading edge for cooling. There is laminar airflow here, but there is plenty of area too. I think this was used on some schneider trophy planes. Very difficult to do for an amateur constructor.

[RFguy]

Been reading quite  alot on approaches to higher performance vehicles, Porches etc, and reducing cooling drag and increasing mass air flow. That's really the name of the game - the mass air flow to  cooling system drag RATIO ..

Most of the recent work has concentrated on reducing interference flow. A large outlet (larger than inlet) can cause interference drag because of the lower momentum / velocity of the exit air compared to the bypass air, and reducing exit airflow area to 50% of the input area yielded  real gains. 

 

The small inlet area , large cowling volume (slows air down, increases pressure)  is not so good unless the exit airflow is handled  . (reduce exit area - but not as a step- must be a transition duct from large back to small. )

 

[skippydiesel]

1 hour ago, RFguy said:

 

SKippy, have you had time to make any progress ?

 

 

 

 

 

Some progress.

 

Son is working, intermittently, on CAD - Hope to have a promising design by end of coming week. Then the whole shebang will be "sectioned" , printed out & assembled as a skeletal female mold. Mold will require a time consuming amount of hand finishing . Probably a month or more befor a cowling can be produced & fitted.

[skippydiesel]

Came across this  EAA Homebuilder Week Sonex video - the relevant (to this  conversation) section commences at about  57 min 10 sec mark

 

https://players.brightcove.net/627008079/7MKHE9SrL_default/index.html?videoId=6293553347001

 

Having two radiators is not new, however the location, well back in the cowling,  is new to me,as is the side air exits.

 

Note;  the speakers comments on front mounted radiators 

[RFguy]

Hi Skip. just watched a few minutes of it.  very interesting ! thanks for posting. Tonight I will sit down and watch the lot.

Where they have the radiators is same as what I am currently considering- ---- where 2 x motorcycle radiators will end up on my Jabiru. That is where there is space !.....

 I am enlarging the front cheeks a little and using 5" SCAT ducts  (or some other suitable 5" ID hose) from the enlarged front inlets back . I have not finished thinking about exit airflow just yet. 

I can 3D print transitions from the SCAT hose to the radiator shape  pretty easily.  (or glass it up) 

 

It's very hard to fit everything at the front of the 230 aircraft because of the cylinder proximity to the inlets (50mm). And the 230 is already nose heavy  so no more nose heavy thanks. 

 

the 125mm inlet size (122cm2 each ) is a bit on the small size so to get the mass air flow and minimise existing cowling disturbance, they need to go right at the front. With only 2 x 122cm2, it might need a third radiator or some other means to get that last 20%. We'll see. 

[skippydiesel]

RFguy - FYI Bunnings Aerospace have a nice line of lightweights, large bore, corrugated hose with a selection of plastic fittings. Vinidex Drain Coil  - 50mm, 65mm, 100mm. The hose itself comes in a non slotted variant. I used  50mm for my two NACA ducts fresh air to panel mounted Mercedes eyeball vents. Worked very well. 

 

You may have to test for suitability in higher temps of cowling.

[RFguy]

yeah  for in cowling  no thanks! that stuff burns with a match . 

 

But you can paint/coat with fire proofing.  There are a few fireproof coatings that are flexible you can paint on/dip...

I have seen this stuff used

 

https://envirograf.com/product/fire-protection-coating-for-pvc-electrical-cables/

 

 

OK for cabin I guess.

[skippydiesel]

What about that foil concertina ducting for kitchens etc - must have a fire rating or at least a high temp resistance.

 

 

 

I suppose you could just splurge & get regular aircraft SCAT, bit of a let down though 😊

 

[RFguy]

The internal surface roughness is a friction agent and it matters. SCAT is not terribly smooth. 5" SCAT is about $100/meter

at 80 kts (full power climb), air pressure at the duct entry, assuming it has clean airflow is 4" water, ~1kPa.

 

assume 125mm inlet (duct size =inlet area)

for 80kts , 125mm ID, roughness of 0mm, 1m long, pressure drop is 110Pa (0.44")

for 80kts , 125mm ID, roughness of 1mm, 1m long, pressure drop is 292Pa (1.17")

for 80kts , 125mm ID, roughness of 2mm, 1m long, pressure drop is 356Pa (1.42")

for 80kts , 125mm ID, roughness of 3mm, 1m long, pressure drop is 400Pa (1.6")

 

So the SCAT will be OK  for that large diameter

for 100mm inlet, 100mm pipe , 1mm rough PD = 390Pa 1.55"

2mm rough PD = 475Pa 1.9"

 

For my application, 125mm scats , one each side, is 250cm2, so a bit down...I need 290cm2, ideally.  I'll see how it goes. I would recommend at least 125mm inlets x 2 for your application.

 

4" inlet x 2 = 156cm2, that is a long way down, I beleive marginal airflow for your application, considering pressurizing the cowling, there will be  leaks and mass air loss.

 

I think you want to assume 1.5 to 2" pressure drop across the radiator with direct hose butted feed  at this air velocity. 

with a transition to use the entire radiator ara (you need to ) It's not so bad- you are going to go from the 125mm dia hose  122cm2  up the (larger) radiator area ~ 300cm2 with a transition manifold, and so the velocity will drop appreciably and so will the pressure drop.  that will allow you to tolerate longer hose, bends etc
 

Edited January 30, 2022 by RFguy

[Geoff_H]

If I remember correctly you can use Reynolds number to calculate friction in pipes etc then use a circuit diagram for the airflow and solve it as if it were am electrical circuit.