Ed,
Good discussion about streamline ducts. No doubt that they are superior although I have a slightly different take on what quality makes them work best. I also agree that it is wall separation that we are trying to avoid.
But IMHO the important way to get there is "avoiding sharp turns." I think of the air molecules as little race cars coming in the duct. The less turning they do, the better. If they need to turn, the radius of the turn needs to be as large as possible. And just as important, the turn radius is distributed so that more of the turn is done after the air has started to slow down (near the face of the radiator). In other words, the turn radius is a function of speed. Just like with a car, don't turn it much before slowing down or it will separate.
With that in mind, see why the "conventional duct" is so terrible. There is a single sharp turn right at the end of the straight-a-way. Separation occurs there and the whole plenum becomes turbulent.
With the bell shaped duct (K&N), it is easy to see why we need length. The longer the duct, the larger the turn radius can be throughout the whole distance.
Given our limited space however, there will undoubtedly be a point where the necessary turn radius becomes too small for the speed of the air and it separates. But at least get the air to expand as much as possible before that happens.
With your restriction in the neck you are setting yourself back before you start the necessary expansion. You have created less distance over which to average the turning radius, you have increased the speed - meaning the air can tolerate less of a turning radius before separating (lower velocities are known to maintain laminar flow much better than high velocity), and you have increased the total amount of expansion the streamlines need to undergo (narrower starting point). So my guess is a rather dismal effect on cooling compared to what you could have.
BUT, since your cooling is still adequate I am sure you have made a very nice overall drag reduction. There is no way a conventional duct with that amount of area would work well. In other works, while I am very skeptical that the restriction actually helped cooling, big kudos to you for absolutely minimizing drag and duct area while maintaining sufficient cooling.
In fact, seeing as how you have proven that it works I am considering doing that for my oil and intercooler ducts as they are currently getting more air than they need...
JMHO
On 2/26/07, Ed Anderson <eanderson@carolina.rr.com> wrote:
Actually, there is, Joe. But, you are going to be sorry you asked {:>).
I spent quite a hit of time studying a tome (Kuchuman and Weber better know as K&W) on air cooling of liquid cooled engines written back in the hey day of high speed mustangs lightenings, spitfires, etc. Sort of the liquid cooling bible. Chapter 12 (the one of most interest to us) showed a duct that reportedly had the best pressure recovery (84% or thereabouts) around for a subsonic duct that they had found. It was called a "StreamLine Duct" (See attached graph - the graph a of the top graph shows the shape of the duct (or at least 1/2 around the center line - sorry for the poor quality).
After quite a bit of studying and thinking about what I had read about cooling ducts, it finally became clear to me that the perhaps top thing that is clearly detrimental to good cooling is having flow separation in the duct. Most of the old drawings of a cooling duct shape followed a sinusoidal shape - rapid expansion right after the opening. It turns out that "traditional" shape is probably one of the worst shapes for a cooling duct (the story why is too long to get into here).
Anyhow, Flow separation leads to eddies and turbulence which casts a "shadow" of turbulent air on the cooling core. Like a shadow, the further away from the core the separation occurs (like near the entrance of the duct) the larger the shadow it casts on the core area. This "shadow" adversely interferes with the flow of air through the core and reduces the effectiveness of the core.
What causes this separation is that as pressure is recovered by the expansion of the duct, the build up of the very pressure recover we want - starts to hinder the boundary layer flow near the wall of the duct. It slows it down and causes it to lose energy and attachment to the duct wall. At a certain point the flow separates and starts to tumble/rotate and depending where (near the duct entrance or near the core) the flow separates, determines how much of the core area is adversely affected. So if the boundary layer's energy level (air speed of its molecules) is maintained at a high level separation is less likely.
So ideally, you would like to prevent any separation during pressure recovery. The Streamline Duct is the so called "Trumpet" duct or "Bell" duct . After the opening, there is a long section of non-expanding duct followed by a rapid expansion into the "bell" shape just before the core. The long non-expanding part of the duct maintains the energy (air flow) of the boundary layer and separation does not occur until well into the "bell" shape expansion.
In fact, it happens way up in the corner of the bell/core interface and affects a very small area of the core.
For full effectiveness the "Streamline duct" from K&W needs a length of 12-17". Well, that's way more distance than I had. So I got to thinking that if keeping the speed of the air molecules near the duct wall helps prevent boundary layer separation and the cooling killing eddy of turbulent air - what could I do with my short 3 - 6" (no jokes you guys). We all know from Bernoulli that if an area is squeezed down that the velocity of the air flow increases - right?
So I decided to try to maintain or increase the energy of the air by pitching down the neck just before it goes into the bell shape expansions in hopes that the increased energy will help the boundary layer stay adhered to the duct wall until well into the corner of the bell shape. So that's the story of the pinched ducts. There is no question in my mind that this is not as effective as if I could have had the 16" to build the duct - but, in this hobby, you work with what you've got - right?
Does it work? Who knows - but I seem to fly with less opening area than most folks and have no cooling problems. So that's my 0.02 on the topic - see told you, you would regret asking {:>).
Ed
----- Original Message -----
Sent: Monday, February 26, 2007 8:53 PM
Subject: [FlyRotary] Re: cowl openings for water radiators
Ed, is there some particular reason that you necked the inlet down small, then enlarged it again. Thankyou for the pictures. JohnD
----- Original Message -----
Sent: Monday, February 26, 2007 3:39 PM
Subject: [FlyRotary] Re: cowl openings for water radiators
John, don't know if these photos will help. But, like you I only have between 3 and 6" of duct distance on the radiators. You should do Ok with 20 sq inch on each opening with a good diffuser/duct. Attached are some photos of my current ducts. The openings are 18 sq inches each. I have had one opening down to as little as 10 square inches - but that was a bit marginal - so opened it back up. I have a generous exit area for the hot air including a larger 4" x 12" bottom opening as well as louvers on each side of the cowl. So you mileage could vary - but Tracy has essentially the same size opening as well as several others.
Ed
----- Original Message -----
Sent: Monday, February 26, 2007 12:12 PM
Subject: [FlyRotary] cowl openings for water radiators
What size openings do I need for the water radiators? The Wittman Tailwind cowl I have has postal slots of 3' x 7 3/4" , which is approx. 22 1/4 sq in. on each side. Sam James for the 160 Lycoming is using 4 3/4' round holes which are
17.6 sq. inches on each side. My radiators are quite close to the opening and I plan on making the diffusers trumpet shaped, will the openings be large enough if I can stay over 20 sq. inches on each side with a decent trumpet shape. JohnD hushpowere II on order - hope to start in 2 weeks if weather cooperates.
--
Homepage: http://www.flyrotary.com/
Archive and UnSub:
http://mail.lancaironline.net/lists/flyrotary/
--
Homepage: http://www.flyrotary.com/
Archive and UnSub:
http://mail.lancaironline.net/lists/flyrotary/
--
Homepage: http://www.flyrotary.com/
Archive and UnSub:
http://mail.lancaironline.net/lists/flyrotary/
|