Cooling Ducts
jeff_spitzer
This is a follow up to the post by Corky Scott regarding radiator
sizing and installations.
I have designed numerous radiator installations for aircraft in my
career and responded once before to a post about inlets. I have
included that post below. This mostly addresses inlet sizing and duct
work. It does not address a couple of other topics in Corky's post
which I think are worth discussing:
Radiator face area:
The rule of thumb noted may be o.k. for a certain type of radiator. I
cannot say either way. What I can say is that all radiators are not
alike and airplanes can take advantage of these differences. I assume
that Jess is referring to a single pass, one layer automotive
radiator. These are normally about 1.5 inches thick. These are great
for slow flying airplanes where only a small pressure differential can
be generated. In a fast flying airplane, much more pressure can be
generated. In this case, stacking radiators, or using thicker cores,
not only saves tons of volume but actually can increase overall heat
exchanger efficiency. This can be done by using a multi-pass type
arrangement where the coolant passes back and forth across the cooling
air with the hottest coolant passing through at the rear and working
it's way forward through subsequent passes. In most cases, two passes
is plenty.
Outlet sizing:
While it is possible for an aircraft with a small speed range to
develop a system with a fixed outlet size, it is not practical for a
typical cross country type airplane. This is because you want to use
high power settings both in climb and in cruise at two very different
airspeeds. If no geometric changes are made, more speed will
necessarily give more flow and more cooling (not to mention the cooler
atmosphere at altitude) and cause over-cooling. The purpose of a cowl
flap is to allow back pressure to be generated at the outlet which
reduces flow, cooling, and drag.
Turning corners:
It is widely accepted that you cannot simultaneously diffuse and turn
in an internal duct. This is too say that it's a very difficult thing
to do without extensive testing and hopefully a little sound theory.
I wasn't sure from the post if that's what was suggested. The post
below describes a way to generate very effective diffusers of nearly
any cross section as long as they are straight. There are exceptions
to the rule such as: with proper inlet conditions you can accomplish
small diffusion values at a very high rate i.e. up to 12+ deg total
included if you're only going about a diameter in length. You can
also diffuse _much_more_rapidly_ as you approach the heat exchanger.
In this case, the restriction is transmitted upstream much like in an
external diffuser and extreme diffusion rates can be had in short
distances. The best duct designs will take advantage of this and I
believe that the P-51 duct is a good example. It's been quite a while
since I've seen a cross section of the P-51 duct but I believe that
the rounded corners and turning and diffusing that Corky was referring
to are all accomplished due to the back pressure effect and would not
be practical without an adjacent flow restriction.
Referring again to Corky's quote of Jess; the 1.23 factor for outlet
to inlet makes no sense to me whatsoever. The duct friction losses
going from point A to point B should be so miniscule in the overall
design so as to be ignored. Therefore, there is no "critical length
ratio" unless perhaps it is all factored into the previous values of
inlet opening, diffuser angle, and outlet opening.
I must admit before I go that I almost always had one advantage at
work that I don't have at home. That is dealing directly with the
radiator manufacturer, often on a customized design, where they
furnish me with heat rejection and pressure drop numbers for their
unit. To go successfully grab a heat exchanger from a junkyard with
no data given is a different story altogether. It is in this area
where more postings are needed as to what type (brand?) of radiator
was used, what is the power output of the engine and what pressure
drop _measured_across_the_radiator that was used. If we could
assemble a couple different operating points for similar radiators we
could put together an extremely useful database for homebuilders where
one could _really_ size a radiator with confidence. Then, supplying
the air/pressure is pretty straightforward.
Happy Cooling,
Jeff Spitzer
On Wed, 17 Sep 1997 20:55:15 -0700, John Burnaby
<jonl...@worldnet.att.net> wrote:
>Jeff,
>
>Thanks for your input. I have saved some articles by Dan Bond over the
>years and there is one thing that is confusing to me. In one article on
>diffuser design, he shows a cross section of a cowling inlet. The
>section shows the cowl lip and then the duct tapering down and then
>expanding at an included angle of about 12 deg.
>
>First, I don't think that this is practically feasible in most
>homebuilts simply because there isn't the distance between the opening
>and the cowl plenum to taper down and then expand at a rate of 12 deg.
>back to the original opening area.
>
>Second, what benefit is gained by constricting the duct, only to open it
>up to where you started. It seems that the dynamic pressure would
>increase through the constriction and then decrease to the value at the
>opening when the diffuser expands. What was gained?
>
>>While most diffusion of the cooling air is accomplished before entering the cowl,<
>
>I don't understand this. Would you elaborate?
>
>I'm building a Glasair, and I think that there is a lot of cooling drag
>that can be eliminated. In that I have to make baffling for the upper
>cowl area anyway, I will probably make plenums for both banks of
>cylinders instead. Six of one......
>
>John
>
I have not seen the article that you are referring to. I can only
guess that the author was suggesting a slight acceleration to align
the flow before entering the diffuser. The air entering the lip will
still be influenced by the direction it came from and the geometry of
the external diffuser (more on that later) making the design of the
internal diffuser difficult. If you accelerate the air slightly after
entering a closed duct, you can straighten it and be more confident in
your diffuser design.
The number you cite: "12 deg included angle" is a good number for a
rectangular cross section with two walls diverging at 12 deg total
included. A better rule of thumb is to always try to maintain about 6
deg total included angle for any cross sectional shape based on an
equivalent (round cross section) conical diffuser. If you work out
the numbers (which I haven't for a long time) I think you will find
that these are very close to the same thing (e.g. 12 deg with 2 walls
vs 6 deg conical) The conical comparison opens up much more
possibilities when you're trying to fit obscure shapes in under your
cowl. I have designed some variable cross section ducts (like
bifurcated inlets) using this and achieved diffuser efficiency >80%.
Also note that when space is constrained and you can only do a little
diffusing, you can push this angle up quite a bit for short distances.
As for the concept of external diffusion, it goes something like this:
Imagine if you will that you have an airplane flying through the air
at 200 MPH with an inlet facing into the airstream. Attached to that
inlet you have your engine intake which consumes a fixed volume of air
based on RPM, MAP and volumetric efficiency. (nit pickers restrain
yourselves :) ) Lets say that with the size inlet that you have
selected that the air that the engine is gulping must move though the
opening at a speed of 100 MPH. I can now tell you how much pressure
you will have at that inlet face. The reason is that, by definition,
you have externally reduced the speed of the air from 200 MPH to 100
MPH. And, because there was no skin friction or turbulence associated
with it (yet), it was 100% efficient. From the point of view of the
pressure recovery of the captured air, what you have done is gone
through a 2:1 area ratio diffuser with 100% recovery. The actual
situation with the finned cylinders is only slightly more complex. In
this case, the back pressure of the cooling flow and capture pressure
will have to be in equilibrium. This may or may not result in enough
cooling flow. Lets say that the back pressure of the system is
"relatively high". In this case, a lot of external diffusion would be
necessary in order to generate sufficient pressure. This is
equivalent to saying that the air entering the cowl will be moving
"relatively slowly". If the air moving through the inlets is moving
slowly then the inlets will have to be "rather large" in order to pass
enough mass/volume of air to effectively cool. If, on the other hand,
the back pressure is "relatively low" then the air can be slowed
(diffused) less and the same mass of air can pass through smaller
inlets. This is good. The reason is that external diffusion is not
"free" as it might sound. Of course there is the same momentum loss
for any type of cooling flow based on change in velocity of the
entering and exiting air mass. With a fully internal diffusion
system, the additional (airframe) losses are rather insignificant.
With external diffusion you have two more drag terms. The first is
the pressure drag of forcing that high pressure inlet through the air
at 200 MPH. The second is the additional skin friction (and possibly
separation) associated with lip spillage. This may be a little too
much for non-aero people without the associated streamline sketches,
but the process of decelerating the air externally necessarily causes
an acceleration of the air passing over the outside of the lip area.
So, which is better internal or external diffusion? As usual, it
depends. If you already have air basically stagnating on the front of
your (front engine airplane, conventional) cowl and accelerating
around the outside anyway, you will not add (lip spillage and
pressure) drag by using external diffusion. In this case it is
basically free. If on the other hand you were to extend the nose to
make room for long internal diffusers and use smaller inlets you would
lose big time. In the case of a rear engined airplane, for instance,
where you don't already have a readily accessible stagnation area, it
will typically be more efficient overall to use internal diffusion if
you already have the internal volume to accomodate it. If you don't
have the volume then a carefully considered compromise is in order.
So, to address the original question. There is very little "inlet
drag" associated with properly designed inlets of the proper size on
the front of a conventional airplane. As such, they can be made
"relatively large" without penalty and the air moving through them
will be going "relatively slowly" As such, there is very little to be
gained by continuing to diffuse the air internally or duct it directly
to the cylinders. Going back to the original post referring to
Nemesis: In this case the airplane has a relatively long slender nose
and tight outside cowl lines. Here he both has the length for
effective internal ductwork and a need for smaller inlets. Plus,
he's racing and every 1/10 percent counts.
Jeff Spitzer
Jeff Spitzer
On 27 Mar 1998 18:10:06 GMT, Charles.K.Scott@**NOSPAM**.dartmouth.edu
(Charles K. Scott) wrote:
>In article <6fgmj2$9...@bgtnsc02.worldnet.att.net>
>"Bruce A. Frank" <BAFRANKMailBlock哦worldnet.att.net> writes:
>
>> As for proper sizing of the radiator, companys such as Griffin and Ron
>> Davis Racing have had years of experience with development of the
>> correctly sized radiator. When you supply them with the parameters of
>> operation, displacement, Horse Power and vehicle speed they can usually
>> draw from their stock to meet our needs. The speeds at which we fly are
>> in the same range as the racing cars for which these companys supply
>> radiators.
>
>Ahhh, this is the kind of information I can really use. Thanks Bruce.
>
>Corky Scott
I have used Ron Davis once in the past and found that he could not
supply me with the information that I needed to my satisfaction. What
I asked was for heat rejection vs. (airflow) pressure drop for a given
core face area. He did not have this. It was more like what Bruce
said where Ron said "...well, this ## x ## radiator cools a Nascar
Grand National car making 500 hp at 150+ MPH..." or something to that
effect. That's good info to have. But it does not answer the
question. The reason why it's so important to know with good accuracy
the required pressure drop is because _that_is_the_only_way_ you can
design a proper inlet and outlet system. The scenario goes something
like this:
-First pick a critical design condition, this is usually hot day climb
at full power. You need to know:
-Indicated Airspeed
-Heat rejection (usually BTU's equivalent to output power)
-Atmospheric conditions
-Next pick airframe locations for inlet and outlet. This is important
because it will be needed before other design parameters can be
addressed. It has been mentioned before "don't put your outlet in a
high pressure area". This is a good idea but shouldn't be a drop dead
factor. The best place for an inlet is right behind the propeller
(duh!) This is because total pressure available there is:
-Free stream total pressure
+Prop induced velocity head
+Prop static pressure rise
A relatively simple propeller analysis formula can give guesstimates
for the second two and I hope everyone knows how to get the first.
-From these chosen locations and conditions we can establish a
pressure differential that we have to work with. I don't have time to
work up a big example so let's just guess that for a 100 mph climb
you get about 6" of water if the inlets are in the normal location and
the exit is at the bottom of the firewall.
-Your next choice is how to use that 6". If your chosen core needs 6"
to reject adequate heat, you have very little choice. You effectively
have to put a inlet lip on the sucker and hang it out the side of the
Cowl. You cannot reduce the outlet size and need 100% efficiency in
the inlet. A better choice is to use a core that uses say 3" of back
pressure. (Note that this is at least twice conventional automotive
practice) Then you have 3" left over to work with to make everything
smaller. In the end, you need to know:
-Pressure differential of cooler
-Air mass flow through cooler for associated pressure drop.
-Next, you set up a tradeoff with where and how you place the cooler.
The common scenario may be to duct air from an inlet on the front of
the cowl. If this duct is small, it may have significant losses from
inlet to cooler. You also have to decide how (where/when) you will
diffuse the air that will pass through the cooler. For instance, if
you cannot possibly squeeze more than a 3" duct around the engine to
the cooler, then there is positively no benefit in making your inlets
any bigger. So, you go through each component step by step keeping a
tally of pressure usage. Things to consider include:
-External diffusion amount
-Inlet loss
-Internal diffuser efficiency and/or effectiveness
-Inlet duct losses
-(core pressure drop)
-Outlet duct losses
-Now you can design your outlet. Your outlet size will be that which
uses up all residual pressure based on previous discussion. Usually,
the outlet will be designed as a "nozzle" and depending on entrance
efficiency will turn substantially all remaining pressure into
velocity. Then, knowing velocity and mass flow leaving the exit, it
can be sized.
-Finally, you rework the numbers considering max level flight speed,
full power. At this condition your heat rejection will be identical
to the climb case but you will have substantially more pressure to
deal with. So, as we all know, you reduce the size of the outlet
until a pressure balance is achieved. i.e. outlet velocity goes up
and area goes down...same outlet mass flow, much increased back
pressure. Note that for same mass flow all other components of
cooling ducting will be identical. So, if you've made it this far
deciding how much cowl flap motion you need is very simple. (BTW,
don't forget about leaks if it applies, you might be surprised how
much flow can go through a loose fitting cowl flange)
In regards to my previous statement about needing data on coolers, it
still stands. If anyone out there has data as to pressure drop and
heat rejection for a type of cooler it would be very useful. In
writing this I realized that I would also have to have duct design
info in order to reverse engineer mass flow from empirical data. Note
that data of heat rejection and flight condition even including inlet
and outlet sizes will not fully characterize a heat exchanger
sufficiently for the above analysis. There must be enough info to get
mass flow. One good way to do this would be to measure temps before
and after the HX. Then from an assumed heat rejection based on power
you could get within about 5% on mass flow based on air temp rise.
Let me say that without good HX design data, the process may be a
little trial and error. But I think the above methodology shows how a
resourceful homebuilder could "develop" a good system.
-Start with some assumptions
-Design an overall system
-Fly and measure pressure differential and temp rise (maybe some other
pressures to if its complex ductwork)
-If it's not cooling adequately, you'll at least know where to work
first and can avoid the common "just make the inlet bigger" scenario.
Charles K. Scott
In article <3516c107....@rosebud.sdsc.edu>
Jeff Spitzer writes:
> Referring again to Corky's quote of Jess; the 1.23 factor for outlet
> to inlet makes no sense to me whatsoever. The duct friction losses
> going from point A to point B should be so miniscule in the overall
> design so as to be ignored. Therefore, there is no "critical length
> ratio" unless perhaps it is all factored into the previous values of
> inlet opening, diffuser angle, and outlet opening.
>
> I must admit before I go that I almost always had one advantage at
> work that I don't have at home. That is dealing directly with the
> radiator manufacturer, often on a customized design, where they
> furnish me with heat rejection and pressure drop numbers for their
> unit. To go successfully grab a heat exchanger from a junkyard with
> no data given is a different story altogether. It is in this area
> where more postings are needed as to what type (brand?) of radiator
> was used, what is the power output of the engine and what pressure
> drop _measured_across_the_radiator that was used. If we could
> assemble a couple different operating points for similar radiators we
> could put together an extremely useful database for homebuilders where
> one could _really_ size a radiator with confidence. Then, supplying
> the air/pressure is pretty straightforward.
I knew this would be a good subject. Jeff, Jess was talking about
using auto air conditioning heat exchangers (sorry I forget what these
are called but they are the under dash units, not the radiator mounted
in front of the engine's coolant radiator). These units appreared to
be rather thick when compared to an auto radiator and compact.
I'm not planning to use something like this anyway because I want to
have a radiator that is specifically designed for the airflow at cruise
for the Christavia. This would be in the 120 to 130 mph range. The
radiator must be compact so that it fits within the cowl area without
requiring large bends in the ductwork. Since the Christavia is such a
big fat airplane, I may have more luck packaging everything in the nose
and keeping the ductwork mostly straight than someone with an RV. I
was planning to eat the cost and have something built by Griffin.
The cooling system for any auto conversion appears to me to be the
single most complicated part of the installation. I can't count how
many stories and articles I've heard about things people have tried
that didn't work, or worked with HUGE penalties in drag.
I was sorely hoping that one could just calculate the necessary fin
area for a radiator based on a formula "cubic inch displacement = X
squared in inches. But as always, this seems to simplistic and doesn't
take into account the things you mentioned like the desired pressure
drop or how thick (front to rear) the radiator should be.
So how about this: Can we list the various inputs one must have in
order to design a cooling system?
I'll start and hope that others can add to the list.
1. Cubic inch displacement
2. Speed range
3. Inlet placement and size
4. Cowl flap
5. Exhaust augmentation
6. Heated air outlet placement
That's all I can think of for starters. Using an auto radiator may be
the least expensive way to find a heat exchanger but if it requires so
many compromises to get it to effectively cool, I'm in favor of the
smaller but more efficient (and more expensive) racing type radiator
either made for the airplane or selected off the shelf that will work.
Corky Scott
Bruce A. Frank
> > to inlet makes no sense to me whatsoever. The duct friction losses
> > going from point A to point B should be so miniscule in the overall
> > design so as to be ignored. Therefore, there is no "critical length
> > ratio" unless perhaps it is all factored into the previous values of
> > inlet opening, diffuser angle, and outlet opening.
> >
> > I must admit before I go that I almost always had one advantage at
> > work that I don't have at home. That is dealing directly with the
> > radiator manufacturer, often on a customized design, where they
> > furnish me with heat rejection and pressure drop numbers for their
> > unit. To go successfully grab a heat exchanger from a junkyard with
> > no data given is a different story altogether. It is in this area
> > where more postings are needed as to what type (brand?) of radiator
> > was used, what is the power output of the engine and what pressure
> > drop _measured_across_the_radiator that was used. If we could
> > assemble a couple different operating points for similar radiators we
> > could put together an extremely useful database for homebuilders where
> > one could _really_ size a radiator with confidence. Then, supplying
> > the air/pressure is pretty straightforward.
>
I sent this info to Corky who thought should post it here. Jerry
Schweitzer has stated in a couple of forum talks that he has not found
it necessary to use a larger radiator exhaust air outlet, that is, his
exhaust air outlet is the same as his cooling air intake. The
requirement is that the exhaust air exits into a low pressure area.
Jerry has built two Ford powered planes, a V-6 STOL and, the plane he
presently flys, an RV-6 and has not had cooling problems with his
method.
As for proper sizing of the radiator, companys such as Griffin and Ron
Davis Racing have had years of experience with development of the
correctly sized radiator. When you supply them with the parameters of
operation, displacement, Horse Power and vehicle speed they can usually
draw from their stock to meet our needs. The speeds at which we fly are
in the same range as the racing cars for which these companys supply
radiators. We do have enough experience with some engines to have a good
handle on the "junk yard" radiators one can salvage for good and
inexpensive performance.
There have also been extensive articles, particularly in a Canadian EAA
publication, on the design of the proper duct work to provide adequate
low drag cooling.
--
Bruce A. Frank, Editor "Ford 3.8L Engine and V-6 STOL
BAF...@worldnet.att.net Homebuilt Aircraft Newsletter"
| Publishing interesting material|
| on all aspects of alternative |
| engines and homebuilt aircraft.|
*------------------------------**----*
\(-o-)/ AIRCRAFT PROJECTS CO.
\___/ Manufacturing parts & pieces
/ \ for homebuilt aircraft,
0 0 TIG welding
While trying to find the time to finish mine.
Charles K. Scott
In article <6fgmj2$9...@bgtnsc02.worldnet.att.net>
"Bruce A. Frank" <BAFRANKMailBlock哦worldnet.att.net> writes:
> As for proper sizing of the radiator, companys such as Griffin and Ron
> Davis Racing have had years of experience with development of the
> correctly sized radiator. When you supply them with the parameters of
> operation, displacement, Horse Power and vehicle speed they can usually
> draw from their stock to meet our needs. The speeds at which we fly are
> in the same range as the racing cars for which these companys supply
> radiators.
Ahhh, this is the kind of information I can really use. Thanks Bruce.
Corky Scott
Charles K. Scott
In article <6fgq1u$mdn$1...@dartvax.dartmouth.edu>
Charles.K.Scott@**NOSPAM**.dartmouth.edu (Charles K. Scott) writes:
> companys such as Griffin and Ron
> > Davis Racing have had years of experience with development of the
> > correctly sized radiator.
Here's some additional info on radiator manufacturers. I tried to find
the Griffin web page but it seems to be down at the moment. But I
found another web site, C&R Racing (http://www.crracing.com/main.html)
and when I called them they were enthusiastic about making radiators
for airplanes and had absolutely no qualms about doing so. They told
me that they had already made some units for several homebuilders.
The woman who answered the phone turned out to be a pilot, although she
hadn't flown for a number of years. Her husband is a contact for the
company and he called me shortly after I hung up.
I described my engine and estimated power and estimated cruise (265
cid, 165 to 170 hp and 120 to 130 mph) and asked if it would be
possible to size a radiator appropriate for those specs. No problem he
said, that's what they do all the time.
We talked for a while about radiators in general and what makes racing
radiators different from street radiators. I found I was under a
misconception. I thought that street radiators had a high fin density
compared to racing radiators but actually it's the opposite. Racing
radiators have a more dense fin spacing because there is more air
passing through them so they can be sized smaller but be more
efficient.
He also said that the intake opening can be surprisingly small to flow
the amount of air necessary to the radiator, as long as you are getting
the necessary amount of airflow through system. This is in regards
racing applications of course, airplanes also have to contend with long
taxis downwind and long holds. I think it's accurate to say that the
cooling system should have the capacity to cool the engine adaquately
for an indefinate period on the ground as well as cool the engine
during high density altitude climbs in hot humid weather.
Interesting stuff.
Corky Scott
Bruce A. Frank
mark.j...@alliedsignal.com
In article <6fgfu9$hp1$1...@dartvax.dartmouth.edu>,> [big snip]
> I was sorely hoping that one could just calculate the necessary fin
> area for a radiator based on a formula "cubic inch displacement = X
> squared in inches. But as always, this seems to simplistic and doesn't
> take into account the things you mentioned like the desired pressure
> drop or how thick (front to rear) the radiator should be.
>
> So how about this: Can we list the various inputs one must have in
> order to design a cooling system?
>
> I'll start and hope that others can add to the list.
>
> 1. Cubic inch displacement
> 2. Speed range
> 3. Inlet placement and size
> 4. Cowl flap
> 5. Exhaust augmentation
> 6. Heated air outlet placement
>
> That's all I can think of for starters. Using an auto radiator may be
> the least expensive way to find a heat exchanger but if it requires so
> many compromises to get it to effectively cool, I'm in favor of the
> smaller but more efficient (and more expensive) racing type radiator
> either made for the airplane or selected off the shelf that will work.
>
> Corky Scott
>
Corky,
A couple of years ago, I did some preliminary design work for a custom built
unlimited racer for a potential sponsor. A major portion of the effort
concerned sizing the cooling system and I was faced with trying to solve some
of the same issues. Although the horsepowers and speeds were a little higher
than what you're probably intersted in, the approach should be the same. I
also collected a number of handy references which you should be able to find
in any good college library.
A good rule of thumb for the cooling requirement is that the heat rejection to
the coolant is equivalent to 25 - 30 percent of the engine's horsepower. I
have some NACA data on heat rejections measurements taken on Merlins and
Allisons at various power conditions and this relationship holds up pretty
well. Take the HP * (0.25 or 0.3) * 0.7069 (conversion factor from HP to
BTU/sec) to determine the required heat rejection capacity. The critical
sizing point is usually where you have high power and low cooling airflow
(i.e.; climb).
Two excellent books to consult are "Aerodynamics of Propulsion" by Kucheman
and Weber and "Compact Heat Excangers" by Kays and London. I would recommend
that anyone interested in cooling design try and obtain copies. Kucheman and
Weber have a good treatment of inlets and exits, including the definition for
constructing a streamline diffuser. Kays and London have extensive
appendices containing test data on the heat transfer and pressure loss
characteristics for a variety of heat exchanger configurations; including
typical radiator cores. If your recollection of heat transfer is a hazy
memory from college, there are some good sample problems worked out which may
be a big help :-)
It is usually not too hard to find a good location for the duct inlet where
the static pressure is high, but locating the exhaust can be tricky. You need
to have a pretty good idea of the pressure distribution. Using the engine
exhaust as a "jet pump" can provide some additional margin and is usually most
effective at low speeds and high power conditions when you most need it. NACA
report 818, "An Experimental Investigation of Rectangular Exhaust-Gas Ejectors
Applicable for Engine Cooling" has a very good treatment on the subject.
I hope this info is helpful. BTW, the racer design was shelved when the
backer couldn't come up with the money (anybody interested?), so I don't know
if my design would perform as predicted. I hope you have better luck.
I also hope nobody is offended by my posting such a long article which didn't
mention Zoom or fried chicken :-0
Mark Johnston
Sr Development Specialist
AlliedSignal Engines
Phx AZ
mark.j...@alliedsignal.com
-----== Posted via Deja News, The Leader in Internet Discussion ==-----
http://www.dejanews.com/ Now offering spam-free web-based newsreading
Bruce A. Frank
Jeff,
It is too bad you did not get the heat transfer data that you desired
from Ron Davis Racing but I did not mean to leave the impression that
most of the these suppliers would talk theory with a builder. What they
will do is provide their experience with similar sized engines,
sometimes the specific engine you wish to use, and what has worked. This
past Fall, Bayard DuPont installed a Ron Racing radiator for the rear
engine of his DEFIANT. Bayard told me that he talked extensively with
them and felt the got more than enough information to sellect the
appropriate design and size from them. If I were inventing a design that
took refined theory development to optimize heat rejection and material
suitability I would go to the books or some consultant in
thermodynamics. I would not expect that a company, which is likely to
sell me only one radiator for my one off project, to hook me up with a
thermo guy on the phone.
In this area I actually want to talk to the hands-on people in the
company, the trial and error race car builders who have found out what
works. I have built many a design right off the engineer's drawing board
that in theory did everything called for, and more, that failed
miserably in application. The calculated approch may get you in the
ballpark but the trial and error guys can save you a lot to time and
money. As Badwatter Bill said a while back, one doesn't have to reinvent
the wheel, get on the phone to the people who have already done it, in
the real world.
I did not intend to lecture here, only to point out that experience may
outweight the theoretical.
--
Bruce A. Frank
mark.j...@alliedsignal.com wrote:
>
> In article <6fgfu9$hp1$1...@dartvax.dartmouth.edu>,
Since your post did not contain anything about chicken or Zoom you are
way off topic, BUT thank you for providing such excellant information
and sources on cooling.
Terry Mortimore
Hi Corky,
This is a timely subject for me. I'm just doing the final assembly on my gear reduction unit.
After the dyno testing it will be on to thinking about prop selection and cooling system
design.
I read the Jess Meyers Contact article with interest as well, do you know what he means by the
statement "By using a tube area of 1 x 125 inches double row, aluminum"?
Does anybody have Ron Davis Racing products snail mail address, I'd like to send down some
information on my installation and see what he suggests for a radiator. I'll do the same for
C&R racing, thanks for their link.
In a past article in Contact there was reference to a man that had experience building
radiators for the Indy cars, I have not had the time to go back through the past issues to find
out who he was but perhaps it was C&R, I see that they are in Indianapolis.
If I can nail down the correct radiator size my P-51 style belly scoop can be designed using
Jess Meyers and Hans Mayers articles.
tailwinds, terry
Terry Mortimore 2.7L Subaru RV-6A
38 Cartier St.
Sault Ste Marie terry.m...@sympatico.ca
Ontario Canada
P6B-3K2 RAA #4061 EAA #229708
Jeff Spitzer
>>In article <3516c107....@rosebud.sdsc.edu>
>>Jeff Spitzer writes:
>>
>>> Referring again to Corky's quote of Jess; the 1.23 factor for outlet
>>> to inlet makes no sense to me whatsoever. The duct friction losses
>>> going from point A to point B should be so miniscule in the overall
>>> design so as to be ignored. Therefore, there is no "critical length
>
>Not true. A streamline diffuser has a pressure drop of 1.08, while a parabolic
>diffuser (and others) has a drop of 1.35. I concider almost 30% greater
>efficiency something that shouldn't be ignored.
>
>---
>David Parrish
Drop of 1.08 vs. 1.35 what? Velocity head...no way. Style points? I
have no idea what you're talking about. If you mean something like
inches of water then surely it would be a function of flow velocity.
My diffusers usually have pressure rises not drops. BTW the snipped
quote referred to the "duct" that connects point A to point B. It was
not meant to include any diffuser if there is one.
Jeff Spitzer
David M Parrish
>> to inlet makes no sense to me whatsoever. The duct friction losses
>> going from point A to point B should be so miniscule in the overall
>> design so as to be ignored. Therefore, there is no "critical length
Not true. A streamline diffuser has a pressure drop of 1.08, while a parabolic
David M Parrish
>A good rule of thumb for the cooling requirement is that the heat rejection to
>the coolant is equivalent to 25 - 30 percent of the engine's horsepower. I
>have some NACA data on heat rejections measurements taken on Merlins and
>Allisons at various power conditions and this relationship holds up pretty
>well. Take the HP * (0.25 or 0.3) * 0.7069 (conversion factor from HP to
>BTU/sec) to determine the required heat rejection capacity. The critical
>sizing point is usually where you have high power and low cooling airflow
>(i.e.; climb).
Here's the calculations I made:
230HP * .38bsfc = 87.4lb/hr (14.57 gal/hr)
Assumes a bsfc equal to best for injected aircraft engine.
Hopefully electronic fuel injection can equal that.
* 18,486 BTU/lb = 1,615,676 BTU/hr
Heat of combustion of gasoline
Better than 1.5 million BTU/hr. Now I see why they call it a
heat engine!
* 25% = 403,919 BTU/hr (118,377 J/sec)
Cooling load (Autos are 17 to 26%)
180 l/min = 3kg/sec
Coolant flow for my engine (@5400RPM)
if Heat = Mass * Specific heat of water * Change in Temperature
118,377 = 3 kq/sec * 4186 J/kgDegC * DeltaT
DeltaT = 9.43degC = 17degF
In other words, for my engine to mantain a temperature at full power, the
temperature drop across the radiator would have to be 17 degrees F.
Also, since anti-freeze has a lower heat capacity and pure water , the more
ethylene glycol that's in the coolant, the higher the temperature drop has to
be.
---
David Parrish
JStricker
David,
Did I miss something here? I understand your reasoning except for the first
assumption you made of 230hp and a .38 bsfc. I'm not sure which aviation
engine you're referring to, but I don't know of any that are making that
kind of economy at close to 100% power. Or were you figuring that for
cruise? I'm a little confused I guess.
I think you need to do it for a worst case scenario which would be a full
power climb for several minutes and then your bsfc will be closer to .45 or
.50 I would guess. Otherwise, your cooling system looks to me like it would
be undersized.
Correct me if I misread your post.
John Stricker
--
Remove the "nosp..........." Oh hell, you folks know what to do and
why I had to put it in. If one of you real humans wants to contact me:
"I didn't spend all these years getting to the top of the food chain
just to become a vegetarian"
David M Parrish wrote in message <351fa...@TmedBSD.MCG.EDU>...
David M Parrish
In article <F43352B0FA2D4EA4.02F7F897...@library-proxy.airnews.net>, "JStricker" <jstr...@odsys.NOSPAM.net> wrote:
>Did I miss something here? I understand your reasoning except for the first
>assumption you made of 230hp and a .38 bsfc. I'm not sure which aviation
>engine you're referring to, but I don't know of any that are making that
>kind of economy at close to 100% power. Or were you figuring that for
>cruise? I'm a little confused I guess.
The engine is a Subaru SVX, 3.3l and rated at 230HP. The 0.38bsfc was a number
I found for best economy for a fuel injected aircraft engine. Don't know where
I found it. If you don't like that number, then use 0.44. That's best
efficiency for an automotive engine from the Hayward book. The SVX engine is a
very good modern engine design with direct electronic fuel injection, so it
should be able to achieve that number.
Also, from the Hayward book, cooling load is 17 to 26% of the heat generated.
I used close to the upper bound for my calculations. With the difference
between 0.44 and 0.38 being only 13.6% more heat, it's still a good ballpark
number.
---
David Parrish
Jacob.O...@jpl.nasa.gov
> The engine is a Subaru SVX, 3.3l and rated at 230HP. The 0.38bsfc was a
number
> I found for best economy for a fuel injected aircraft engine. Don't know
where
> I found it. If you don't like that number, then use 0.44. That's best
> efficiency for an automotive engine from the Hayward book. The SVX engine is
a
> very good modern engine design with direct electronic fuel injection, so it
> should be able to achieve that number.
>
> Also, from the Hayward book, cooling load is 17 to 26% of the heat
generated.
> I used close to the upper bound for my calculations. With the difference
> between 0.44 and 0.38 being only 13.6% more heat, it's still a good ballpark
> number.
>
> ---
> David Parrish
>
I also have a Subaru 3.3 liter. But I would like to get more power out of it.
Does anyone know of any tuners out there that have experience with the engine
?
Thanks
Jake
Nigel Field
JStricker wrote:
>
> David,
>
> Did I miss something here? I understand your reasoning except for the first
> assumption you made of 230hp and a .38 bsfc. I'm not sure which aviation
> engine you're referring to, but I don't know of any that are making that
> kind of economy at close to 100% power. Or were you figuring that for
> cruise? I'm a little confused I guess.
>
He was refering to the Subaru EG-33 high performance 6 cylinder 10:1
CR car engine which indeed makes 230 HP and a BSFC of .38 when run closed
loop EFI.
Calculating the engine heat output is a good first start but then it gets
real messy with lots of variables that cannot be defined. Little things
like inlet lip shape, boundary layer, outlet location, and many more can
make a big difference, very hard to quantify. I found that starting with
a good theoretical solution and adding 50% then iterate on the airplane
till you get it where you want it is the most practical solution. Mine
took 3 re-works till I was satisfied with cooling verses drag
performance. Still I have it right on the ragged edge such that down low
sustained full power on a 30 deg C day and it will boil. So I just
watch it carefully. Above 4K in cooler thinner air, less power developed
less heat, no problem for good cruise. But I hate drag and would rather
have 5 more kts than bullet proof draggy cooling as it was when I
started.
Nigel Field EA-81 Vari-eze
JStricker
David,
I think that was probably the Continental as used in the Piper Malibu, with
aggressive leaning. I seem to remember something about it being the most
fuel efficient engine available at the time. If it was, that was at (I
think) 65% power. The fuel flows got up to almost 25 gal/hr at 100% (at
least on the one I flew back when they first came out) which is somewhere
around .48 bsfc. Greg Travis probably has a much better memory than I do
about it.
Just curious, have you plugged this into a spreadsheet to plot some curves?
I think your formula is interesting and some could argue about the Bsfc's
all day and everybody be right.
My interest was in how things came out and how you intend to size the
system. For instance, under certain conditions, most all new cars will
overheat. Low speed, hot day, high load. The cooling system just isn't
designed for that extreme. Now that's not to say it won't work most of the
time, but sometimes it won't be adequate.
You see that on a/c engine as well. Often I have to level for a bit to keep
temps as low as I'd like on one of our 108 degree days after a bit of
waiting/taxiing to take off. That doesn't mean there is a deficiency in the
system, just that it is operating in conditions outside of its design
parameters.
If you set the formula up on a spreadsheet, then loaded in varying
airflow's, bsfc's, etc., you could find out what the critical condition was.
It may be on the ground, at idle. It may be on high power/moderate airspeed
climb. Who knows. It's an interesting formula though. And you also have
to get good bsfc's for the engine from somewhere. For a/c engines, that's
not too hard, but for auto conversions, it's a little harder unless you want
to log some serious dyno time.
I also wonder how the cooling load is affected by block/head materials. Is
it less for aluminum components, as they can radiate more on their own? I'd
think so, but I'm not sure. Bruce or Greg---any ideas?
John Stricker
--
Remove the "nosp..........." Oh hell, you folks know what to do and
why I had to put it in. If one of you real humans wants to contact me:
"I didn't spend all these years getting to the top of the food chain
just to become a vegetarian"
David M Parrish wrote in message <3520e...@TmedBSD.MCG.EDU>...
>In article
<F43352B0FA2D4EA4.02F7F897...@library-proxy.airnews.ne
t>, "JStricker" <jstr...@odsys.NOSPAM.net> wrote:
>>Did I miss something here? I understand your reasoning except for the
first
>>assumption you made of 230hp and a .38 bsfc. I'm not sure which aviation
>>engine you're referring to, but I don't know of any that are making that
>>kind of economy at close to 100% power. Or were you figuring that for
>>cruise? I'm a little confused I guess.
>
JStricker
Nigel,
The EG-33 is rated (for your use or whatever standard you prefer) 100% at
230hp with a bsfc or .38?? Or is that a cruise bsfc at a lower power level?
I don't keep up with the soobs at all, what displacement is that? Enquiring
minds want to know.
No, seriously. I've done quite a few dyno runs and I've NEVER seen bsfc's
this low at full power. Never. Unless, it's capable of much greater power
than that and this is, in effect, a lower power setting.
I whole-heartedly agree with your cooling system design philosophy though.
It would be foolish and unnecessarily draggy to make it stay cool in even
the most extreme conditions indefinitely. Just design it to take care of
your engine under normal (plus a safety margin of course) conditions and
realize that you can't always climb at 100% power indefinitely on hot days.
John Stricker
--
Remove the "nosp..........." Oh hell, you folks know what to do and
why I had to put it in. If one of you real humans wants to contact me:
"I didn't spend all these years getting to the top of the food chain
just to become a vegetarian"
Nigel Field wrote in message <35217E...@pwgsc.gc.ca>...
>JStricker wrote:
>>
>> David,
>>
>> Did I miss something here? I understand your reasoning except for the
first
>> assumption you made of 230hp and a .38 bsfc. I'm not sure which aviation
>> engine you're referring to, but I don't know of any that are making that
>> kind of economy at close to 100% power. Or were you figuring that for
>> cruise? I'm a little confused I guess.
>>
>
David M Parrish
>I also have a Subaru 3.3 liter. But I would like to get more power out of it.
>Does anyone know of any tuners out there that have experience with the engine
>?
I think Reiner Hoffman did some work with the 3.3, though he never got it to
production. (Not sure why.) I think he was getting 260+ out of it.
---
David Parrish
David M Parrish
In article <57FC3F23A3D732B1.F978D229...@library-proxy.airnews.net>, "JStricker" <jstr...@odsys.NOSPAM.net> wrote:
>least on the one I flew back when they first came out) which is somewhere
>around .48 bsfc. Greg Travis probably has a much better memory than I do
Hayward's book on internal combustion engines says 0.44 is best for an
automotive engine, yet Jhonny said he was getting 0.40 on a dyno'ed Legacy
2.2l at 175HP. Go figure, though I should get a better bsfc than most
Lycosauri.
>Just curious, have you plugged this into a spreadsheet to plot some curves?
Not yet. My next step is to talk to Griffin Radiators, then plug back into the
references I have. I've even thought of doing some experiments with two of the
most popular junkyard radiators - the GM airconditioning evaporator and the VW
Diesel Rabbit radiator.
>My interest was in how things came out and how you intend to size the
>system. For instance, under certain conditions, most all new cars will
>overheat. Low speed, hot day, high load. The cooling system just isn't
>designed for that extreme. Now that's not to say it won't work most of the
>time, but sometimes it won't be adequate.
I'm still in the reseach phase, so no hard numbers or fiberglass ductwork.
I do have had some ideas about low speed cooling without the drag penalty of
an oversized system. Since mine is a pusher, I can have the radiator exhaust
ported to just infront of the prop. (If you have this big fan, you might as
well use it.) Having the exhaust pipes terminate inside the radiator duct
might also give a bit more suction. And if all else fails, there's always the
possibility of a spraybar in front of the radiator. Evaporating water can suck
up a tremendous amounts of heat.
---
David Parrish
David M Parrish
>Drop of 1.08 vs. 1.35 what? Velocity head...no way. Style points? I
>have no idea what you're talking about. If you mean something like
The number is deltaP/deltaPsubzero. DeltaPsubzero is the pressure drop across
the radiator, deltaP is the pressure drop across the diffuser and radiator. So
a streamline diffuser looses only 8% of the airflow energy compared to the
radiator alone, compared to 35% for a short sine diffuser. That 27% boils down
to more drag. (The pressure drop equals more work to get the air through the
system and larger radiators to compensate for the reduced airflow through the
system. Both equal more drag.)
Clear?
---
David Parrish
Yoram Leshinski
Is that the same engine that Crossflow aero have? They claim 320HP heigh
output turbo charged, 250HP standard.
Yoram Leshinski
Jeff Spitzer
Your response is very clear. I've never lumped the diffuser and
radiator together like that before as a system and that seems to be
why we see things differently. I approach each element of the system
separately. What I understand from your post is that a "streamline"
diffuser will be the most efficient type when followed immediately by
a heat exchanger. I fully agree with this though I'm not sure any of
us can actually design one of these with pencil and paper.
Going back to the original post, however, I still don't believe that
anything that has been said here relates to this magic ratio of 1.23
for outlet length to inlet length as quoted. I meant to say: the
length of inlet and outlet duct is a rather insignificant feature
compared to the inlet design, diffuser design, outlet design etc. So,
where is the magic in duct length ratio?
Jeff Spitzer
On Mon, 06 Apr 1998 13:17:16 GMT, d...@TmedBSD.MCG.EDU (David M
Parrish) wrote:
>I'm not good enough with the math either, but I'm using numbers from a table in
>_Aerodynamics of Propulsion_ by Kuchemann and Weber. I can post the table
>if anyone wishes.
>
>---
>David Parrish
I would be very interested in this table. I don't have the reference
you cite and though I have seen many streamline ducts, I have never
actually seen an equation or table for one. Hopefully the chart gives
values as a function of head loss ratio or something that relates HX
pressure drop to flow head. It seems to me that the ideal shape would
vary accordingly where a HX with high pressure drop could have an
extremely shaped duct and one with no pressure drop would have zero
shape allowed.
Jeff Spitzer
Jeff Spitzer
On 6 Apr 1998 19:54:04 GMT, Charles.K.Scott@**NOSPAM**.dartmouth.edu
(Charles K. Scott) wrote:
>In article <35253475....@rosebud.sdsc.edu>
>Jeffrey...@nospam.gat.com (Jeff Spitzer) writes:
>
>> It seems to me that the ideal shape would
>> vary accordingly where a HX with high pressure drop could have an
>> extremely shaped duct and one with no pressure drop would have zero
>> shape allowed.
>
>Jeff, when you say "extremely shaped duct" may I assume you mean the
>inlet is a small size compared to the size of the chamber leading to
>the heat exchanger?
Yes, sort of. In my terminology an inlet is the thing that first
receives air from the free stream. It may be any shape, size,
whatever and may or may not play any role in diffusion.
A "streamline diffuser", (at least what I think David Parrish is
referring to), represents a diffuser which is necessarily followed
immediately by a significant flow restriction such as a heat
exchanger. This downstream restriction allows for diffusion rates
which would not be possible without it. The shape will be influenced
by the ratio of "entrance" head (not to use the term inlet) to the
head loss across the restriction. An "extremely shaped duct" may have
an area ratio of 10:1 in a length of not more than a few diameters of
the entrance size. This is something that cannot be done without the
restriction. If we examine why it will actually help explain
external diffusers as well.
If you imagine a duct of say 3" diameter with air flowing in it at 200
ft/sec. This is about 588 CFM of air and has a dynamic head of about
9 in. of water. Lets say the radiator is 10" in diameter and has a
static pressure drop of 3 in. of water at 600 CFM (when introduced
uniformly!). For the given face area of .54 sq. ft. then the uniform
face velocity is about 18 ft/sec. If I put my 3" duct very close to
the heat exchanger i.e. use a short or "extremely shaped duct", it
won't be ideal. What will happen is that my 9 in. of dynamic head
will impact the radiator and force a disproportionate amount of the
flow through the impact area. After all, the radiator only had about
3" of drop at 18 ft/sec. Typically, the ratio of frictional losses to
turbulent losses through a heat exchanger will make the pressure drop
approximately proportional to face velocity^1.5. So, my 9" will get
me a little more than twice as much flow velocity through the impact
area. If for simplicity we call the impact area 3" dia. then about
100 CFM (17% of 600) goes through what amounts to 9% of the face area.
That's the non-ideal part. The good news is that the other 500 CFM
has to find some other way to get through. So, it spreads out and
goes through (horribly over simplified) at about 17 ft/s with plenty
of excess static pressure (which we'd like to hold on to make the rest
of our system smaller, particularly the outlet). Good and bad- Good:
I passed all the air through the heat exchanger and still had some
static pressure left over. Bad: I wasted a whole lot of static
pressure in the impact area by passing the air through the HX too
fast. I certainly did not pick up enough heat rejection benefit
(although more heat was rejected locally) to make it worth it. Now
comes the fun part: Knowing that I have to go through a 11:1 area
ratio and knowing that I don't want to have this impact loss, how
close can I put the 3" duct to the 10" radiator and what should the
duct shape be? In actuality, one of the gross simplifications that I
made in my example was to ignore the upstream pressure generated by
the impact. What that does is tell the oncoming air to start moving
to the side well before it hits the HX. So, the impact area would
only be 3" if the duct were within about 3" from the HX. As I pull it
back and make the "shape less extreme" the upstream pressure has
longer to work spreading out the impact area until it's a non issue.
Hence, a streamline diffuser. What's happening is you are simply
shaping the walls in a natural way for the flow to spread based on the
downstream restriction. I'm still waiting for someone to post a
method to describe this shape for various conditions. Before I
finish, let me change the scenario. Let's say that the HX has a
static pressure drop of 18" and that the same 3", 588 CFM is
introduced. Obviously, there needs to be some additional static
pressure available from somewhere e.g. outlet suction or
pressurization at the inlet to our 3" duct. But if we go back to the
diffuser then you can have a very extremely shaped duct, almost
absurd. The reason should be obvious. The 9" of dynamic head will
only raise the flow velocity passing through the impact area by about
30% vs. 200% for the previous case. So, in my over simplified
example, you can pretty well jam the little 3" duct right up to the
face of the 10" radiator and see very little additional loss compared
to a long diffuser. In fact, (what I think David Parrish was
referring to), you may actually have less loss than a long diffuser
due to the greatly reduced skin friction loss in the short diffuser.
>
>By the way, could you re explain what it means to have an externally
>diffused duct as opposed to an internally diffused duct?
By an internally diffused duct, I mean a shaped duct section with
increasing cross section.
An externally diffused duct is an oxymoron. External diffusion refers
to that which takes place before entering the inlet in the free
stream. In the example above, all the air could not move through the
impact area with the dynamic head that it had. So, it had to spread
out and find another way through. When it spread out it naturally
diffused itself. The same thing happens as the air approaches the
inlet. The pressure that exists at the face of the inlet is an
equilibrium condition between the flow approaching it and the
downstream restrictions. Let's say the velocity approaching the inlet
is 200 ft/s (or about 9" water) If the total of the downstream
restrictions requires about 4" of water then the air will have to
diffuse prior to entering the inlet. It will naturally do this as I
said. If you trace the streamlines which attach to the upper and
lower (lets talk 2-D) lip of your inlet you will find that at some
upstream point they will be parallel (and closer together than the
inlet width) and represent the free stream. As they approach the
inlet they naturally are spread apart by the downstream restriction.
This is your sorta free external diffuser. In fact, if you trace the
streamlines you will find the area ratio between lip and "capture
area" to be about 1.5 (in my example) which gives the required
pressure rise. The not free part is that as these streamlines diverge
(slow) in the capture zone, they converge (accelerate) in the adjacent
areas. This makes the velocity passing over the lip and cowl higher
than if the inlet were not there or were not diffusing. If instead we
had made the inlet .66 the size that it was and diffused our 1.5 area
ratio inside the cowl then the streamlines would have entered
substantially parallel to the free stream and would not cause
additional drag. In that case we would have to make room for the
internal diffuser and possibly have additional losses carrying the
faster moving air from the inlet to the HX.
>
>Thanks, Corky (struggling to understand) Scott
David M Parrish
>a heat exchanger. I fully agree with this though I'm not sure any of
>us can actually design one of these with pencil and paper.
I'm not good enough with the math either, but I'm using numbers from a table in
_Aerodynamics of Propulsion_ by Kuchemann and Weber. I can post the table
if anyone wishes.
>Going back to the original post, however, I still don't believe that
>anything that has been said here relates to this magic ratio of 1.23
>for outlet length to inlet length as quoted. I meant to say: the
Not sure where that one came from. (Not me.) Most of what I've read has to do
with inlet and outlet duct shapes and opening areas.
---
David Parrish
Charles K. Scott
In article <35218462....@rosebud.sdsc.edu>,
Jeffrey...@nospam.gat.com wrote:
> >Going back to the original post, however, I still don't believe that
> >anything that has been said here relates to this magic ratio of 1.23
> >for outlet length to inlet length as quoted. I meant to say: the
> In article <3528d...@TmedBSD.MCG.EDU>
d...@TmedBSD.MCG.EDU (David M Parrish) writes:
> Not sure where that one came from. (Not me.) Most of what I've read has to do
> with inlet and outlet duct shapes and opening areas.
The duct ratio of 1.23 inlet to exhaust was specified by Jesse Meyers
in his article on cooling systems for auto engined homebuilts in the
last issue of Contact Magazine.
He stated that it was something he found to be necessary in his
experience. He did not say why.
Corky Scott
Charles K. Scott
> It seems to me that the ideal shape would
> vary accordingly where a HX with high pressure drop could have an
> extremely shaped duct and one with no pressure drop would have zero
> shape allowed.
Jeff, when you say "extremely shaped duct" may I assume you mean the
inlet is a small size compared to the size of the chamber leading to
the heat exchanger?
By the way, could you re explain what it means to have an externally
diffused duct as opposed to an internally diffused duct?
Thanks, Corky (struggling to understand) Scott
David M Parrish
>I would be very interested in this table. I don't have the reference
>you cite and though I have seen many streamline ducts, I have never
>actually seen an equation or table for one. Hopefully the chart gives
>values as a function of head loss ratio or something that relates HX
>pressure drop to flow head. It seems to me that the ideal shape would
>vary accordingly where a HX with high pressure drop could have an
>extremely shaped duct and one with no pressure drop would have zero
>shape allowed.
0.00 | 1.000 1.000 1.000 1.000
-0.10 | 0.769 0.797 0.847 0.886
-0.20 | 0.638 0.680 0.757 0.812
-0.30 | 0.561 0.608 0.694 0.761
-0.40 | 0.505 0.554 0.648 0.724
-0.50 | 0.463 0.515 0.613 0.694
-0.60 | 0.431 0.484 0.586 0.671
-0.70 | 0.405 0.459 0.563 0.651
-0.80 | 0.383 0.438 0.544 0.634
-0.90 | 0.365 0.420 0.526 0.618
-1.00 | 0.350 0.404 0.510 0.604
-1.20 | 0.328 0.384 0.489 0.585
-1.40 | 0.311 0.367 0.472 0.568
-1.60 | 0.297 0.353 0.456 0.553
-1.80 | 0.285 0.340 0.442 0.540
-2.00 | 0.275 0.329 0.430 0.529
-2.20 | 0.267 0.320 0.420 0.520
-2.40 | 0.260 0.312 0.412 0.512
-2.60 | 0.255 0.306 0.406 0.506
-2.80 | 0.252 0.302 0.402 0.502
-3.00 | 0.250 0.300 0.400 0.500
The first column is the X dimension relative to the radiator height, so a 12"
high radiator would have a 18" duct. (12"/2*3) The next columns are for 25%,
30%, 40% and 50% opening to radiator area ratios. Each column is the Y
dimension relative to the radiator height. The result looks like the bell of a
musical horn.
Which you choose depends on the application and radiator. The ratio of
radiator to inlet area is inversely proportional to pressure and directly to
velocity, so a 50% ratio would double the pressure at the radiator and halve
the velocity of the incoming air. 25% would be 4x pressure, 1/4 the velocity.
Below about 1/3 area ratio, things start getting critical and care has to be
taken to prevent flow seperation. (Turning vanes can help here.)
---
David Parrish