Mr. Meridith's Mysterious Effect

Discussion in 'Mazda Rotary' started by Monty, Oct 23, 2010.

Help Support HomeBuiltAirplanes Forum by donating:

  1. Oct 23, 2010 #1

    Monty

    Monty

    Monty

    Well-Known Member

    Joined:
    Jul 15, 2010
    Messages:
    1,283
    Likes Received:
    478
    Location:
    Fayetteville, AR / USA
    Well by request I'm posting this. Can't find the original spread sheet just now, but I found the charts.

    Explanation to follow.
     

    Attached Files:

  2. Oct 23, 2010 #2

    Monty

    Monty

    Monty

    Well-Known Member

    Joined:
    Jul 15, 2010
    Messages:
    1,283
    Likes Received:
    478
    Location:
    Fayetteville, AR / USA
    What I did to arrive at this was take an engine I had the heat balance on and do an analysis just like the cooling system was a ram jet. Instead of a combustor I put the heat from the coolant into the airflow. I then mixed the engine and radiator exhaust. I assumed a heat exchanger effectivness of .8. All procesess were ideal with no losses. I may have assumed nozzle and diffuser efficiencies of .9. I can't really remember. Then I expanded the flow to atmospheric and calculated the thrust. Thrust is converted to hp using forward speed. I converted to HP, because there is no airframe. This was just to give an idea of what you could get if everything was perfect. I took compressibility into account.

    Lednicer is the guy to listen to on this. Far more of an expert than I am.

    I found this googling a bit.

    The authoritative source for liquid cooling is Kucheman and Webber (sp?). I have been trying to get my hands on a copy forever.

    I also like Jack D. Mattingly's book, though it is very specific to Jet engines.

    Thermodynamics of Propulsion by Hill and Peterson. which is more general and easier to adapt to odd problems.
     
  3. Oct 24, 2010 #3

    Monty

    Monty

    Monty

    Well-Known Member

    Joined:
    Jul 15, 2010
    Messages:
    1,283
    Likes Received:
    478
    Location:
    Fayetteville, AR / USA
    It occurred to me that these charts may be misleading.

    My point in doing this analysis was to try to save people a lot of wasted effort. I constantly run into people who want to try to create thrust from their cooling system by using the heat from the exhaust. While cooling drag reduction is a worthy goal, trying to make a subsonic waste heat ram jet is a waste of time. If you have very high true airspeed and are making over 1000 hp, some meager thrust may be possible, but it is highly unlikely. The best you can do is perfection, and even if you achieve that the gains are small. In the real world, you will never get perfection.

    Cooling drag is caused by momentum loss from the air taken on board. Think of scooping up lake water in a bucket from a speeding boat. The force you feel is drag. If you can throw the water back out of the boat in the aft direction faster than you are going, you get thrust. If you can't you get drag.

    So we should try to minimize the amount of cooling air we take onboard, because you are never going to be able to make thrust. It's just a matter of making it less bad.

    Interestingly the Reno unlimiteds employ some unique twists on this concept. They under-size the radiator (less cooling air=less drag) and dump ADI fluid on the rad to make up the cooling difference. A phase change makes for huge heat transfer and a large increase in volume. Now you CAN make thrust. Another sacred cow is that you must have a boundary layer diversion scheme. Boundary layer diversion is only to make sure the diffuser operates at max efficiency, or to have uniform flow into a jet engine or fan. To do this produces a large drag penalty. Early on diffusers were designed like a reverse nozzle. The flow is very sensitive due to the adverse pressure gradient. It will become detached and turbulent very easily. So you need a long very gradual diffuser and very good inlet conditions. Eventually it was discovered that you can use the obstruction of the radiator to your advantage. The streamline diffuser was developed and if you really must know the gory details, Kucheman and Webber have the answer. The streamline diffuser can be much shorter and is less likely to suffer from separation. If you use boundary layer air for cooling, you get a double benefit. One, you don't have to pay the drag penalty for the diversion scheme, and two you are re-accelerating air that you have already paid the drag penalty on! If you don't believe me, just check out the scoop on Strega. They have given up on the ram jet idea and instead are ingesting lower energy boundary layer. You get less pressure recovery at the radiator as a result, but who cares!! You re accelerate it by expanding the volume with ADI fluid.
     
    Last edited: Oct 24, 2010
    Synergy likes this.
  4. Oct 24, 2010 #4

    Starman

    Starman

    Starman

    Well-Known Member

    Joined:
    May 17, 2009
    Messages:
    2,011
    Likes Received:
    61
    Location:
    High in the Andes Mountains
    Neat! I looked up ADI fluid and found this very interesting website: More on cooling airplane engines** Very neat information there and what you have.



    This means, technically, that they are turning racers into canards (lifting tail/rear wing), it's marginal but it's there.
     
  5. Oct 25, 2010 #5

    autoreply

    autoreply

    autoreply

    Moderator

    Joined:
    Jul 8, 2009
    Messages:
    10,732
    Likes Received:
    2,542
    Location:
    Rotterdam, Netherlands
    Firefox just quitted me, now I can re-type the whole sory :(
    Do I understand correctly that you took the exhaust heat into account as well? (Since that's the majority of the engine heat)
    I've never fully understood the need for it. First of all, in our region (up to 250 mph or so), boundary layers are thin and thus, momentum loss is pretty small. Also, like you said, you're re-energizing that same boundary layer again. Thirdly, having a "lip" to not ingest the boundary layer has a lot of extra area, producing drag as well.

    Your comments about a ramjet got me thinking. I found a nice tool of NASA, doing all this automatically ;-)

    EngineSim 1.7a beta

    So, what did I do?
    Input:
    Speed; 250 mph, sealevel, Inlet recovery, 0.9, and otherwise, everything default.

    Then, we've got the inlet/outlet ratio. I set it to 1.
    [​IMG]
    Then we can fiddle around with the inlet area, until the fuel flow seems correct. I used the fuel flow to "tune" it, because that's exactly what defines HP. Since about 60% of the fuel is transferred to heat, you take 60% of your total fuel flow. Note that many engines (especially the conti/Lyc) are running rich and thus producing less heat than you'd expect. Hence I use the 100HP Rotax 912, which burns 20 liters/h @ 75%. So that's 33 lbs/hr for 75HP.

    An inlet area of 0.1 sqft gives this fuel consumption. It produces a net thrust of 177 lbs and has an (internal?) drag of 45 lbs. That's 131 lbs, or 600N.

    250 mph is 112 m/s. 600N @ 112 m/s is 67 kw, or almost 90HP. 60% of this is 40kw, or 54HP of net thrust. Total thermal energy was (0.6/0.4)*75HP=112.5 HP, so the "thermal propulsion" is only 50% efficient.

    So what goes wrong in my calculations/assumptions?

    I suspect, the "exchange efficiency" of the heat is very low, because of the very small temperature differences.
     

    Attached Files:

  6. Oct 26, 2010 #6

    Monty

    Monty

    Monty

    Well-Known Member

    Joined:
    Jul 15, 2010
    Messages:
    1,283
    Likes Received:
    478
    Location:
    Fayetteville, AR / USA
    You can't just put the energy as a fuel flow into the tube. The key is the heat rejection temp of 100C not the adiabatic flame temp! Of course you get thrust that way!

    You have to do a heat balance and figure out how much air you need to cool at CpDeltaT, and you have to use heat exchanger effectiveness to know how hot you can get the discharge air. There is no way you will get choked flow at the nozzle either.

    For the exhaust you do uniform reversible mixing with the radiator discharge. (something you will NEVER get in practice) Use EGT and engine air flow expanded to conditions aft of the radiator. The amount of cooling air is large, and the temp is 100C or so (IF you do a really good job), by the time you mix the two streams, the engine exhaust is like peeing in the Amazon. You would be better off just pointing the exhaust aft through a nozzle to atmospheric and using the energy for thrust.
     
    Last edited: Oct 26, 2010
  7. Jun 11, 2013 #7

    Aircar

    Aircar

    Aircar

    Banned

    Joined:
    Feb 20, 2010
    Messages:
    3,567
    Likes Received:
    367
    Location:
    Melbourne Australia
    This is an interesting subject --experts still argue over the relative contribution of the (thick) laminar profile wing versus the cooling system 'negative' drag on the Mustang as compared to the Spitfire (thin non laminar higher critical mach no smaller etc ) but also having a Meredith radiator . Peter Garrison spent a lot of effort on his air cooled ducting and variable area outlets on the Melmoth 1 (and two also I suspect ) and wrote about the possibilities of recovering some lost power --including the 'ejector exhaust' effect which probably occurrs too far down the system to add any useful thermodynamic effects but should allow for smaller inlets (he also used the overwing low pressure area to exhaust into for further air extraction as did his mentor John Thorp. I tried to get some synergy with the Opal by taking air in at the wing root stagnation point and exiting near static on the tailcone with the exhaust via the same opening --many years later I used a pair of VW Golf radiators (four altogether) in tandem to cool the bob Jane Monza racecar --with two radiators in series the air has two passes to pick up heat (the hottest radiator aft for maximum air/glycol differential) qnd the physival dimensions of the cooling ducts are minimized along with contraction and expansion angles --in the case of the Monza the air was exited at the lower corner of the "A" pillar but the windscreen was also replaced by a polycarbonate "marguard" hard surface substitute that curved smoothly inside the former sharp corner where the windscreen and sideglass met (so it was like the exhaust exiting over the cambered part of a wing and hence 'sucked' out . Some photos were posted kindly by Head in the clouds on a much earlier thread. It seemed to work - 640 HP (dyno tested not theoretical) never overheated and in fact on one race (which it won) someone else's decals covered half of one duct (the ducts split around the engine block ) and it still didn't cook itself. Incidentally I was amazed to see how they bed in' a new race engine --to give the effect of a few hundred k's of running without the time and to stop polished bores ---"Ajax" industrial cleaners was fed down the intake trumpets sans air cleaners --the exhaust was psychedelic...
     
  8. Jun 11, 2013 #8

    Norman

    Norman

    Norman

    Well-Known Member

    Joined:
    Nov 29, 2003
    Messages:
    2,880
    Likes Received:
    919
    Location:
    Grand Junction, Colorado
    It, the boundary layer gutter, was required on later models of the P-51 because at about 400 mph the fuselage BL buffeted the oil cooler and created a rumble that caused metal fatigue in the aft fuselage of the "A" model.
     
  9. Jul 11, 2013 #9

    yellow_submarin

    yellow_submarin

    yellow_submarin

    Well-Known Member

    Joined:
    May 10, 2013
    Messages:
    88
    Likes Received:
    2
    Location:
    Laval, France
    I would be carefull with the ramjet comparison. As far as I'm concerned, my domain is limited to subsonic, incompressible (M<0.3) airflow. Thus, the cooling duct should be a divergent to compress air, not a convergent as a ramjet. I might be stating the obvious but I felt the need for it. I can't use the EngineSim applet on this computer, how do you set the compression ratio of the ramjet ?
    There is a compromise to be made : if you choose to slower the radiator air flow, the drag will be decreased but pressure hence temperature will increase compared to a faster flow speed (talking about identical mass flow obvisouly). So as stand-off, the delta T between radiator and flow is decreased thus cooling potential is decreased, which means less thrust.
     
  10. Jul 11, 2013 #10

    rv6ejguy

    rv6ejguy

    rv6ejguy

    Well-Known Member

    Joined:
    Jun 26, 2012
    Messages:
    3,749
    Likes Received:
    2,792
    Location:
    Calgary, Alberta, Canada
    I agree, the comparison with a ramjet is essentially invalid as we are working with far lower expansion rates and temperatures.

    If I understand what you are saying, you are concerned about the temperature rise of the cooling air during diffusion in the duct? In practice this is only a few inches of water at best so this temperature rise is almost immeasurable. Radiator exit air temperature would be almost identical at the same mass flow. Diffusion would however slow the air at the radiator face allowing more time in contact with the radiator fins, thus increasing the radiator exit air temperature and the potential for expansion and therefore thrust when we converge the flow downstream of the radiator. My flying experiments show that you are unlikely to ever get net thrust from from a radiator setup, only offset some drag through good design. To date, I am showing a 85% recovery in momentum at about 120 knots. Further instrumented flight testing to come.
     
  11. Jul 11, 2013 #11

    Monty

    Monty

    Monty

    Well-Known Member

    Joined:
    Jul 15, 2010
    Messages:
    1,283
    Likes Received:
    478
    Location:
    Fayetteville, AR / USA
    The ramjet example is completely valid.

    Thermodynamically it doesn't matter what the mach number is. The calculations are based on the mach number. The hardware is independent from the calculations. It only matters what the pressure rise, heat transfer, and efficiency numbers are. I used absolute best case scenario....and you will never get that. Even doing so the amount of recoverable energy is minimal. The heat rejection temp is just too low to be useful.

    What you can't do is take something like an online calculator that has all the wrong equations and compare that to our situation. I'm sure that the NASA applet uses several assumptions :

    1. choked flow at the nozzle
    2. adiabatic flame temp in the combustion chamber.

    Neither assumption is applicable and will result in thrust numbers that are way too high.
     

Share This Page



arrow_white