[Stoves] Controlling the Primary / Secondary Air Split in ND-TLUDs

Julien Winter winter.julien at gmail.com
Fri Jan 29 16:28:02 CST 2016


Hi Crispin and all;

In an earlier posting on “Riser Height and a 'Counter-Current' Woodgas
Burner - YouTube Vid”:

http://lists.bioenergylists.org/pipermail/stoves_lists.bioenergylists.org/2016-January/011227.html

Crispin said:

“...

The split between primary and secondary is uncontrolled, for all intents
and purposes, and should not be. In the same manner that you advocate at
the end that reducing the area of holes through which the primary air can
enter, this should be applied to the secondary air as well, arranged in
such a way that the incoming has to go to one or the other. It should not
be controlling one and then the other is free to flow more or less
depending on the riser height. In short, don't use a macro structure as a
fine tuning air control mechanism.

There are several products that work in that 'controlled split' manner
though they are hard to find. The point is to decide in advance what the
pri/sec split should be, create it, test it to be sure it is correct, then
scale the draft accordingly. If you were to conduct this test on a SeTAR
type real time test bench, it would provide a very useful set of excess air
and CO/CO2 ratio charts the quickly identify the ideal conditions for the
pri/sec split. This will be independent of the riser height. In such a case
the riser height will affect the power, not the split, then the power can
be regulated by the gross control of the air entering. This is a different
approach taken be nearly all the TLUD designers so far which is to regulate
the primary and let the secondary run free according to flame temperature
and architecture.

If you have control over the entry and the correct split by construction,
the riser height is used for total available draft, and for directing that
energy (a risen has 'power') to the correct mixing of the air and gas, not
the flow of secondary air.

...”

Having a single user control valve for air entry into a stove, then having
the design of the determine the proportion of secondary/primary air is an
interesting idea.  I can see how this could be obtained in part by the
relative magnitude of buoyancy along the path of secondary air vs. primary
air, with the possibility of some constricting holes to help guide the
split to the best ratio.

Clearly, the method you propose has had some success.  However, I still
wonder how a mechanism could handle a few challenges.

===============

1) Turning down a ND-TLUD is very sensitive to small amounts of primary
air, because it doesn’t take very much primary air to support a flaming
pyrolytic front in a small fuel pieces like pellets or nut shells.   Would
it be difficult to get precise, and responsive control with a general air
regulator?

2) As the stove is turned up, the proportional demand for oxygen in the
reactor increases and the gas burner decreases.  Tom Reed reported for the
“Turbo Stove” burning pellets that the ratio secondary:primary air changes
from 6.2 to 3.1   (Reed, TB; Anselmo, E; Kircher, K.  2000.  Testing and
modeling the wood-gas turbo stove.  Presented at the Progresss in
Thermochemical Biomass Conversion Conference, Sept. 17-22, 2000, Tyrol,
Austria).   Can changes in relative magnitude of buoyancy in the gas burner
vs. reactor be used to determine this?

3) Although #2, above talks of a forced draft TLUD, we can still get large
differences in the requirement for primary depending on the fuel we are
burning.   Wood chips are more reactive than pellets, even though they are
the same thickness.

The big effect, I believe comes as the thickness of fuel particles goes
from 5 mm pellets and chips to 3-5 cm thick pieces of wood.  The nature of
the gasification reactions changes, and increase the demand for oxygen.
Pellets produce an ideal migratory pyrolytic front.   That is because the
heat generated from partial oxidation of pyrolytic gases is sufficient to
pyrolyze the centre of the pellets.   The thickness of the migratory
pyrolytic front is not much bigger than a pellet.  However, as fuel
thickness increases, it takes longer for the centre of pieces of wood to
reach pyrolysis temperatures.  The outside of a piece of wood can be 800 °C
when the centre is still around ambient temperatures, because wood, and
especially char, are not great conductors of heat.  As a consequence, more
primary air is needed to generate more heat to complete pyrolysis.  That
means more combustion of char, which demands a much higher flow of air than
woodgas.

In short, when the fuel is pellets we have an ideal TLUD; when the fuel is
large pieces of wood, the TLUD reactor becomes closer to a typical
combustion stove, having char and woodgas combustion occurring side by side.

Some of what I am talking about can be seen in graph I have attached, from
an experiment I conducted in 2014.  The reactor temperature range for
pellets was between 550 and 750 °C, so there was not too much char
combustion happening there.  However for 3-5 cm pieces of wood, the reactor
temperature was between 800 and 1200 °C; those are char combustion
temperatures.  The diagram shows that the fuels used in my experiment fall
into to groups (1) pellets and wood chips, and (2) large pieces of wood.

Char combustion generates a lot of heat, so that would increase buoyancy to
draw primary air.   The question is, would a single air regulator and
primary/secondary air splitter be able to accommodate the changes in
combustion reactions was we go from 5 mm to 5 cm fuel?

4) There are differences in resistance to gas flow in the fuel beds
composed of different materials.  That means that the more of the buoyant
force generated in one fuel bed could be expended in overcoming friction
than in another fuel bed.

5) Sometimes the fire in the reactor has reached the grate well before all
the fuel above has pyrolyzed.  This happens ignition front follows channels
in the fuel bed, or has reached the grate via falling embers.   With
woodchips, unless there are big spaces between the chips, the ignition
front will almost always channel down the sidewalls of the reactor.   The
resultant woodgas is hard to ignite, and we get a lot of smoke.

The operator may be able to resolve the smoke problem by a large increase
in primary air.  Could this be expeditiously done by a single air regulator
and a primary/secondary air split?

===============

So there are the challenges.   I suspect that part of the solution is in
having ND-TLUD stoves designed for specific fuels; namely, (a) pellets and
nut shells,  or (b) pieces of wood.   This is a reasonable thing to do,
because the reactor for pieces has to be wider and twice as long as for
pellets.  Not many people like to wait around for 2-3 hours while 30 cm of
pellets burn.

Cheers,
Julien.

-- 
Julien Winter
Cobourg, ON, CANADA
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.bioenergylists.org/pipermail/stoves_lists.bioenergylists.org/attachments/20160129/06b524df/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Temperature_vs_Char_ALL 01.png
Type: image/png
Size: 134870 bytes
Desc: not available
URL: <http://lists.bioenergylists.org/pipermail/stoves_lists.bioenergylists.org/attachments/20160129/06b524df/attachment.png>


More information about the Stoves mailing list