X-Mozilla-Keys:                                                                                 
Fredmoreno@aol.com wrote:

A common fallacy is that one should attempt to get the maximum water
temperature drop to get maximum heat transfer.  Some even slow the water flow
with a restriction to get this effect.  Not true.

The maximum heat transfer occurs when the temperature drop times the coolant
flow rate is maximized, and this occurs when heat transfer is maximized, and
this is maximized when the radiator surface temperature is as hot as possible
providing the maximum difference between fin surface temperature and air
temperature.  This in turn would occur at an infinite coolant flow rate with
near zero temperature drop but with all the surface at the maximum
temperature.  So as far as the radiator is concerned, maximum flow rate is
good, and the higher the flow rate the smaller the radiator can be.

(When people report improved cooling because of the addition of restriction
to "slow the flow down" what is usually happening is that the pressure drop
was increased, and local boiling caused by low pressure was reduced.  Wrong
perceptions are thus perpetuated.)

It turns out that all heat exchanger performance curves (cross flow, counter
flow, parallel flow) all collapse to the same curve for the case when the
coolant side capacity rate (flow rate in pounds per minute times the specific
heat) is very large compared to the air side capacity rate.  For radiators,
the coolant capacity rate is typically 5 or more times the air capacity rate,
and so this condition is approached quite closely.  So at high flow rates, it
does not matter what the configuration is: lots of coolant means the cross
flow radiator works just as well as the counter flow design. The temperature
of the fins and coolant is about the same all over.

In my analysis I found that the EG/water mixture impedes heat transfer
substantially on the coolant side.  If you combine this with a low flow rate,
you get a high boundary layer resistance on the coolant side that reduces
radiator performance.  You need to keep the flow rate up not just to keep the
temperature difference (coolant to air) across the radiator maximized, but
also to reduce this resistance.

In a car operating usually at low speeds (RPM) the film resistance can be
appreciable. The US radiator approach increases the velocity in the tubes by
a factor of three by folding the flow back and fourth into 3 passes in the
same size radiator.  The price is higher pressure drop, but radiator pressure
drop is usually so low that it does not impede flow appreciably.

In our planes, the engines are turning up at high RPM all the time so flow
rates should be high.  However, because we are going for maximum heat
transfer (smallest possible radiator) it helps to have the flow velocities
high in the radiator.  My analysis shows that for a radiator using 40% E.G.
in water and using 0.1 inch by 1.0 inch oval coolant tubes (usual radaitor
design), you want a minimum of 0.5 gpm per tube, and 0.7 gpm per tube would
be nice.

So count tubes, multiply by 0.7 to get the ideal flow rate.  If it is a lot
more than the water pump rate, then you can consider a larger (or faster)
pump, or folding the coolant passages as in the US radiator design to get the
low coolant side resistance you seek.   This is one reason why 30 gpm for a
180 HP rotary is not enough.

Fred Moreno

The Aircraft Rotary Engine Newsletter.              Powered by Linux.

http://home.earthlink.net/~rotaryeng/            http://www.linux.org