|
|
Hi, Ed. I'll embed a comment or two below - in early part of your analysis.
David
----- Original Message -----
From: "Ed Anderson" <eanderson@carolina.rr.com>
To: "Rotary motors in aircraft" <flyrotary@lancaironline.net>
Sent: Sunday, April 03, 2005 7:32 PM
Subject: [FlyRotary] Re: Cooling -Learned a lot
> Dave, one point made in the stuff I have read on diffusers is that at our
> speeds the air density is considered constant - no meaningful increase in
> air density occurs. I also believed at one time that was the reason we
got
> more cooling with lower velocity (I mean it makes sense, greater density =
> more mass to carry away the heat). But having been disabused of that idea
> from the material I have read, I now have a different understanding of
> what's happening.
Ed, I acknowledge you've read a LOT more than me (I'm being lazy and letting
your do the heavy lifting!!!)
However, "considered constant - no meaningful increase in air density
occurs" in diffusers, I'd take that with a grain of salt - actually with a
grain of "show me the numbers". "Meaningful" to one guy with a particular
view may be "not meaningful" to another with a different perspective. I'd
have to see the formulae to be able to comment further.
> While there is no meaningful density change in a diffuser (subsonic
> velocities), there IS a pressure increase as Tracy pointed out.
Increased
> pressure can result from either an increase in density (which we are told
> does not happen to any meaningful extend in diffusers of interest) and/or
an
> increase in temperature (again, no significant variation of temps in the
> diffuser) OR an increase in the air molecules momentum mV (mass *
Velocity).
> It is the latter that appears to happen in a diffuser, the average
momentum
> of the air molecules in the diffuser is increased by the high energy
> airstream entering the diffuser and therefore average velocity of the air
> molecules in the diffuser being increased (since the mass of the molecule
> does not change).
I don't think I follow the above: "OR an increase in the air molecules
momentum mV . . . It is the latter that appears to happen in a diffuser, the
average momentum of the air molecules in the diffuser is increased by the
high energy airstream entering the diffuser and therefore average velocity
of the air moliceules in the diffuser being increased. . ."
- First, I think we have to consider the air molecules at the face of
the diffuser (widest part of the diffuser) are going to be slower than any
molecules up front, coming into the diverging duct (no convergent section of
the duct - only the "back half" of a venturi tube). So I don't see how
their momentum and velocity can be "increased". Velocity 'increases' in the
first half of a full venturi tube, not in the aft half.
. . . . .-- I think we need to visualize how the molecules are slowing as
they spread out into an increasing duct cross-sectional area - with the
pressure is increasing above static at duct opening (would have to account
for "external diffusion" increasing "static pressure" at opening of duct
being higher than "static pressure" in free air away from the aircraft).
. . . . .-- Visualize packets or "2 dimensional planes" of 500,000
molecules entering the duct at 120mph (176 ft/sec) every hundredth second,
i,e, a spacing of 1.76 inches, then, ignoring wall friction and viscosity,
those "planes of 500,000" will be closer together (fore-aft or axially) at
maybe 1.56 inches (wild guess for discussion) - thus higher density - it
increases as we go aft toward front face of radiator.
. . . . -- The number of molecules entering the duct per unit of time is
constant. The number of molecules at any distance aft of duct opening (at
any plane) is constant (ignoring wall friction and turbulence). Pressure is
increasing as we go aft with speed decreasing (per Mr. Bernouli) - pressure
must be above static atmospheric - significantly, I'll bet. I think that
equates to increased "density" (successive "planes" are closer together.
The "planes" are nothing but an attempt at a description of a calculus-type
approach to a "tiny number for 'time' " or "many iterations of air packets
entering duct per unit of time". Actually, in a given "plane"
(time-distance slice) the molecules would be spreading farther apart to fill
the increasng cross sectional area (getting less dense radially) but because
of the slowing and packing of "planes" the "axial density" is increasing.
The net effect of "spreading out radially" and "packing together axially" is
????? More dense?.
> Since in effect these air molecules are in a fixed space, their higher
> momentum (which is in the form of an increase in air molecule velocity)
Therefore, with momentum being Mass X Velocity, and with V going down, I
think momentum of an individual molecule decreases - but there are
increasing molecules per cubic inch as we approach the radiator at the back
of the diverging duct. I don't see 'momentum' being a 'measure of merit' -
mass flow at a "lower" speed is the measure of merit in extracting heat from
the radiator fins at the lowest drag of those fins.
> results in an increase in the average number of impacts with the core
walls
> per unit time. Higher velocity means the molecule transverse the same
> distance in less time resulting in more impacts per unit time. It is
these
> contacts and the resulting transfer of heat energy from the core
fins/tubes
> to these air molecules that is the major heat transfer mechanism as I
> understand it (radiation being a very minor contributor at these temps).
I think what follows must be re-evaluated with real numbers instead of
"engineering approximations" and "engineering rules of thumb" that use terms
like "meaningful". That's my view of what those little molecules are
doing.
. . . . . David.
> Mass flow (p*Area*Velocity) is indeed fixed and does not change once the
> flow begins in the duct/core system. Since the density is considered
> constant that leaves only the area and velocity as variables. Their
product
> (A*V) must be also be a constant through the system A1*V1 = A2*V2 and you
> would have the same density air at both locations A1 or A2 position. But,
> the pressure of the air (constant density) may indeed be different and if
> greater at position A2 (core passage) than A1 (say duct inlet) then
greater
> heat transfer would occur at position A2 due to the higher average number
of
> molecules impacting the walls of the containment. This even though the
mass
> flow is the same at both locations. Greater pressure =>more air molecule
> contacts with core fins/unit time => more heat transfer/unit time. Of
> course, if you have no mass flow velocity. then the same molecules would
be
> involved time after time and would very soon be saturated with all the
heat
> they can carry. But, since we have those molecules constantly replaced
with
> fresh molecules, heat continues to be transferred from core fins/walls to
> the air.
>
> At least that is what I think I understood. Bill, we need you here
>
> Ed A
>
> ----- Original Message -----
> From: "David Carter" <dcarter@datarecall.net>
> To: "Rotary motors in aircraft" <flyrotary@lancaironline.net>
> Sent: Sunday, April 03, 2005 5:42 PM
> Subject: [FlyRotary] Re: Cooling -Learned a lot
>
>
> > I wonder - is it not more correct to say: Behind the diffuser, the
> velocity
> > will be slower, density higher, therefore there will be "nearly the same
> > mass flow of air" thru the radiator, with same cooling, BUT "at slower
> > speed, higher air density, and therefore less cooling drag"?
> > - Drag is a function of velocity squared. The air density factor
is
> > not squared, thus we seek a reduction of drag by cutting velocity thru
the
> > rad.
> >
> > Bernie, you are the SR-71 PW engine air duct man - am I even close to
> > expressing any useful and true info above?
> >
> > David Carter
> >
> > ----- Original Message -----
> > From: "Ed Anderson" <eanderson@carolina.rr.com>
> > To: "Rotary motors in aircraft" <flyrotary@lancaironline.net>
> > Sent: Sunday, April 03, 2005 4:15 PM
> > Subject: [FlyRotary] Re: Cooling -Learned a lot
> >
> >
> > You are absolutely correct, Tracy.
> >
> > I did not make it clear but the diffuser does the velocity reduction and
> > increases the pressure in front of the core by recovery of (some)
dynamic
> > pressure component of the air flow. This higher pressure in front of
the
> > core then results in an increased pressure differential across the
core.
> > This increase in pressure differential across the core, as you stated,
> > actually speeds up the air flow through the core itself.
> >
> > My apologies for being less than careful on that point.,
> >
> > Ed A.
> > ----- Original Message -----
> > From: Tracy Crook
> > To: Rotary motors in aircraft
> > Sent: Sunday, April 03, 2005 12:24 PM
> > Subject: [FlyRotary] Re: Cooling -Learned a lot
> >
> >
> > Excellent summary Ed, correlates with my experience as well. Only
> > exception I would take is in the following excerpt:
> >
> > "A good diffuser will reduce airflow
> > velocity through the core which will reduces cooling drag. Pressure
> > across
> > the core is increased which further enhances cooling."
> >
> > A good diffuser will reduce velocity but the reduction occurs IN the
> > diffuser, not through the core. As counter-intuitive as it may sound,
> the
> > velocity through the core is HIGHER than it would have been without the
> > diffuser's velocity decrease (and pressure increase).
> >
> > Think about it this way, How could velocity through the core be
reduced
> > by a pressure increase? It isn't. The velocity at this point (through
> the
> > core) is increased.
> >
> > This is the single most misunderstood detail in liquid cooled engine
> > systems.
> >
> > Tracy
> >
> >
> >
> > Subject: [FlyRotary] Cooling -Learned a lot
> >
> >
> > Too right, Jerry
> >
> > My first 40 hours or so were in the marginal cooling zone. {:>).
As
> > other
> > things in this hobby, there are so many variables that interact,
that
> > what
> > may appear simply at first, is almost always a bit more complex. I
> > say(Cooling Axiom 1) if you have enough cooling surface area and air
> > mass
> > flow then it WILL cool. However, you may incur a high penalty in
> > cooling
> > drag - which may not be as important for draggy airframes (such as
> > biplanes)
> > as it is to sleeker airframes. Also a system that adequately cools
> an
> > engine producing 150 HP may not cool an engine producing 180 HP.
> > Picking
> > your cooling design point is important. Optimizing for cruise and
> your
> > will
> > be less than optimum for take and climb. Optimize for climb and you
> > will
> > probably have more cooling drag than required at cruise.
Compromise,
> > compromise - cowl flaps are sometimes used to try to have the best
of
> > both
> > worlds.
> >
> > Some folks advocate a thinner, larger surface area core -which is
> great
> > for
> > slow moving automobiles stuck in traffic with low dynamic pressure
> > potential, but I think is not the optimum for most aircraft. Once
you
> > trip
> > the airflow and turn it turbulent you have incurred most of the drag
> > penalty. Larger surface area cores disrupt a larger airstream and
> incur
> > more drag. Yes, thicker cores produce a bit more drag than the SAME
> > frontal
> > area thinner cores. But, with a thicker core you can use a core
with
> > smaller frontal area.
> >
> > The NASCAR radiator's average 3" thick and on the long tracks
where
> > speeds
> > are higher some even go up to 7" thick. My contention is their
> > operating
> > environment is more akin to ours than regular automobiles moving at
> > slower
> > speeds. You know that the NASCAR folks will spend $$ for just a
tiny
> > advantage - so clearly they don't use thick cores because it is a
> > disadvantage. But, some folks will continue to point to the large
thin
> > radiators designed for environments with much lower dynamic pressure
> as
> > being the way to go. Will it cool? sure it will (Cooling axiom 1
> > above).
> > Is it the lowest drag option for an aircraft of the RV/TailWind
type,
> I
> > am
> > convinced it is not.
> >
> > The diffuser makes a considerable amount of difference and can made
> the
> > difference between a system that cools adequately and one which does
> > not.
> > The biggest culprit that lessens cooling effectiveness is turbulent
> > eddies
> > that form inside the duct due to flow detachment from the walls.
> These
> > eddies in effect act to block effective airflow through part of the
> > core.
> > So keeping the airflow attached to the sides of the diffusers is
> crucial
> > for
> > good cooling from two standpoints. A good diffuser will reduce
airflow
> > velocity through the core which will reduces cooling drag. Pressure
> > across
> > the core is increased which further enhances cooling.
> >
> > I have gone from a total of 48 sq inches opening (total) for my two
GM
> > cores
> > and that provided marginal cooling - down to 28 sq inches (total)
with
> > adequate cooling with an engine now producing more HP.
Experimenting
> > with
> > the diffuser shape made the difference.
> >
> > The K&W book (Chapter 12) really provided the insight to how and
which
> > diffuser shapes provided the better dynamic recovery. The
Streamline
> > duct
> > was shown to be able to provide up to 82% recovery of the dynamic
> > pressure.
> > Some folks reading the chapter misinterpreted the chart to show only
> 42%
> > recovery where there chart was actually only showing the pressure
> > recovery
> > contribution due to the duct walls and did not include the
> contribution
> > due
> > to the core. On the same chart, an equation (which apparently gets
> > ignored)
> > clearly shows that the TOTAL pressure recovery is 82%.
> >
> > I have taken the Streamline duct as a starting point, but since I do
> not
> > have the space to provide the 12-14" for a proper Streamline duct, I
> did
> > some "creative" things to try to insure that there was no separation
> > even
> > though my walls diverge more rapidly than the Streamline duct.
Won't
> > claim
> > mine are as good as a Streamline, but they clearly are much better
> than
> > the
> > previous design which basically just captured the air and forced it
> > through
> > the cores.
> >
> > FWIW
> >
> > Ed Anderson
> > RV-6A N494BW 275 Rotary Hours (Plugs Up)
> > Matthews, NC
> > eanderson@carolina.rr.com
> >
> >
> > ----- Original Message -----
> > From: "Jerry Hey" <jerryhey@earthlink.net>
> > To: "Rotary motors in aircraft" <flyrotary@lancaironline.net>
> > Sent: Sunday, April 03, 2005 9:27 AM
> > Subject: [FlyRotary] Re: phase I flight restrictions was:N19VX flys
> >
> >
> > > It was not long ago that "cooling" was the major issue. Now it
> seems
> > > that we have learned enough to make several different
configurations
> > > work. I can't lay my finger on what it is we have learned but my
> > > recommendation is to use smaller radiators and EWPs. Jerry
> > >
> > >
> > >
> >
> >
> >
> >
> > >> Homepage: http://www.flyrotary.com/
> > >> Archive: http://lancaironline.net/lists/flyrotary/List.html
> >
> >
> >
> >
> > >> Homepage: http://www.flyrotary.com/
> > >> Archive: http://lancaironline.net/lists/flyrotary/List.html
>
>
>
> >> Homepage: http://www.flyrotary.com/
> >> Archive: http://lancaironline.net/lists/flyrotary/List.html
>
>
|
|