[Stoves] Heat transfer and in-line water heater

[Crispin Pemberton-Pigott]

Dear Andrew

Thanks for the thoughtful post.  This discussion is getting to the heart(s)
of the matter.

>> On a shiny pot (heat transfer efficiency) might be greatly reduced 
>> (another 50%?)

>Who knows? The little infra red thermometer I use assumes E=0.95 and 
>in practise a painted radiator doesn't read much different from a 
>copper pipe.

Shiny aluminum is about 0.35 to 0.4 if it is not polished.  This is a severe
reduction, agreed?  The heat is mostly reflected back to the fire raising
the temperature of the char and probably drops the CO a bit.  It is really
complicated...

>OK but it looks like we are looking at 20-30% heat transfer into the 
>pot, that looks like quite a lot of scope for improvement, 

This is the core of the issue.  The skirts, wind protection, black pots,
'right' gap and so on are all aimed at improving it.  Decent heat exchangers
are FAR better than most pots.

>if we're wasting it it must be measurable as temperature of flue gas above 
>the pot.

Therein lies a major possible error.  The losses are not measurable by the
temperature of the exhaust air alone.

Consider the case where the exit temperature is 250 deg with excess air at
100%.  Then consider a case where the exit temperature is 210 degrees, but
the excess air is 3 times higher at 300%.

Which is the more efficient?  The heat transfer efficiency of the higher
temperature will be better of course, but which is losing more heat? Unless
you include the volume of gases passing under the pot, you don't know the
loss, so you don't know the thermal efficiency.

Reducing the gas speed through the mechanism of reducing the excess air
usually sees an increase in heat transfer efficiency and an increase in the
exit temperature _if_ the gas path is short and the system efficiency low (a
single pot stove).  If the gas path length is long and the system efficiency
high (a fish dryer), the same change will also usually see an increase in
efficiency but a drop in the exit temperature. Interesting, eh?

>> Increasing the gap between the flame and the pot will reduce the
>> rate of heat transfer in the centre of the pot and allow more even
>> heating of the whole bottom be allowing the edges to pick up more
>> of it.

>Or by allowing more flue gas re circulation?

Well... that is if it allows for a greater air flow without burning more
fuel (higher excess air).  Since changing any dimension changes the fuel
burn rate and the excess air, it is pretty difficult to know what the cause
of a performance increase is.

>...what we do know is that it can be modelled as a poor 
>conductor in series with the metal of the pot.

Agreed, in most cases, and apparently not all.  I reviewed the issue in
detail some time last year I think in this forum and gave references so no
need to rehash it.  Two years ago Peter Scott, Reg Barrow and I had a
detailed conversation about the matter.  No one could come up with any proof
that air, even one molecule thick, is literally stationary against the
surface of a pot.  It is a useful metaphor, and useful in modelling, but the
idea that still air 1mm or so thick clings to a pot surface is one of faith,
not fact.  The efforts taken to remove or 'deal with it' produce results
because of factors such as dwell time, or most importantly, excess air and
fuel burn rate changes.  Dale Andreatta's work on skirts (I love his
experiments!) showed that the skirt-pot gap didn't matter much at all until
the skirt was a restriction on air flow and then, assuming a high EA, making
a small gap made a  big difference to thermal efficiency by cutting EA and
increasing the flame temperature and reducing EA losses.  I believe that a
great number of experiments done on Rocket Stoves at Aprovecho also support
this anaysis.

Regards
Crispin