[Stoves] [ethos] Re: Laboratory Comparison of the Global-Warming Potential of Six Categories of Biomass Cooking Stoves

[Crispin Pemberton-Pigott]

Dear Friends

Given the stress caused by the disagreements around testing methods I agree
that it would be good to push forward the agenda proposed early this year at
the ETHOS conference.

The main issues have not been debated in terms of what they are.  Here are
five:

Issue 1.	The fuel is not representative of field conditions.  It is
extremely dry and does not burn like regular fuel.  There seems to be no
reason given by Aprovecho for using fuel this dry.  No one presents
justifications for it.  Yet over the strenuous objections of people
including myself, tests that purport to 'compare' stoves are still being
made and the results published, at least on the internet, sometimes in
documentary form.

Why does it matter?  
The fuel used in these tests cannot be used to establish performance
benchmarks.

Stoves that burn sticks on the end, like the Rocket  Stove, perform
extremely well using very dry fuel. Other stoves including the Vesto, Juntos
gasifiers, semi-gasifiers and any stove with a significant preheating of air
do not do well because the fuel burns far too quickly, often running out of
air, producing high levels of CO and particulates.  The Rocket Stove is
optimised to burn that fuel; the stoves it is compared with are not.

Issue 2.	The use of lids on pots for stove tests.  This has been
discussed many times.  Stoves that are designed to operate at a certain
power level are apparently tested with no lids on the pots, forcing the
operator to run the stove above its design power output in order to boil the
water in a reasonable time. This affects emissions.  During simmering, more
fuel than normal is required to maintain the water temperature, even though
the stove may be optimised to simmer at a very low power level.  The Rocket
Stove is optimised to operate a power level to boil pots without lids -
years of research reports show this.  While some concession was made this
year with regard to lids, it is not yet clear to me that the tests recently
published were done according to the stove design parameters, or according
to the UCB-WBT.

Why does it matter?
Unrepresentative tests give unrepresentative results.  Very important fund
decisions are made on the basis of test results, such as the level of CO2
offset funding a stove type is going to receive.  If the test is
unrealistic, the funder and manufacturer are being compromised.

Issue 3.	Fuel heat rating.  There is a portion of the WBT that
calculates the heat yielded during a test.  This means that taking into
consideration the heat in dry fuel, the moisture content of that fuel and
the char remaining, an amount of heat offered to the pot is calculated.  The
UCB-WBT method is very accurate for only one species of wood and one
charcoal type, but quite inaccurate for different wood species, different
fuels, high moisture levels and different charcoal heat contents.  

Why does it matter?
Tests done at different locations are difficult to compare with the ones
done at Aprovecho unless the actual lab data is made available so the tests
can be recalculated.  This data, in general, has not been made available by
Aprovecho so the tests cannot be re-evaluated, nor can the claims made for
comparative efficiency or fuel consumption be verified.  The results from
several months of testing different fuel moisture levels are not rendered
acceptable because so much money and effort was put into them.  They become
acceptable when the protocol is realistic and the calculations made as
accurately as possible.

Issue 4.	Thermal efficiency calculation method.  The formula used to
calculate the efficiency of a stove involves comparing the heat offered to
the pot with the heat absorbed by the pot.  The one used in the UCB-WBT is
not accurate.  There are errors introduced by the lack of lids: no account
is taken of radiation from the hot upper water surface and the excessive
evaporation with no lid.  The bigger error is the one that starts in the
fuel heat rating and is then transferred to the efficiency calculation.  As
the amount of charcoal remaining increases, as the heat content of the fuel
decreases and as the moisture content of the fuel rises, the error gets
larger and larger.  For gasifiers and users of low heat content fuels, the
error can be well over 100%.  By this I mean that the calculated heat
released can be double the actual value, giving the impression that the
thermal efficiency of the stove is only 1/2 what it really is.  In other
cases, the calculated value is very accurate.  Such calculation errors
render meaningless comparisons made between stoves using wood of different
species and moisture contents and, for example, rice hulls, as has been done
with the latest set of published tests.

The forthcoming test results for higher levels of fuel moisture mentioned by
Dean are probably going to be presented within the UCB-WBT framework.  If
so, the thermal efficiency calculations and specific fuel consumption
figures are invalid because of deficiencies in the methodology and formulae.
The only immediate solution would be that the raw data for the tests be
published so that the real performance can be calculated using various other
methods, though that would still not solve the fuel moisture problem.  

Why does it matter?
The UCB-WBT in its present form cannot be used to establish realistic
performance benchmarks.

Issue 5.	Variable and incompetent operation of stoves being tested.
It is always a problem to get people who do not know how to use a stove
really well to test it.  The problem will not easily go away.  The best
answer I have seen is for developers to run their own stoves using fuel of
their choice.  To do that at a central facility is very expensive.  Testing
a different places in the world and comparing notes requires an agreed
protocol.

I have suffered many times with products (other than stoves) being badly run
by 'independent' testers so the problem is not unique to stoves.  In the
case of the recent rice hull stove evaluation, the tester ran out of fuel
part way through the test, yet reports the result which a reader must assume
as being representative of the stove's performance.  This is not reasonable.
Previous tests showed the CO emission to be 1/7 of this test.  This is
inconsistent and frankly, unbelievable.

Why does it matter?
It is dangerous because unfair accusations of bias can emerge if a stove is
underperforming at Aprovecho when being tested alongside other expertly run
stoves and the comparative results published as representative of the stoves
performance in general.

Improving the situation:
A very large amount of goodwill have been shown to the people and
institution of Aprovecho by many skilled and experienced people. No one will
benefit from the publication of results that are correctly critiqued
followed by more of the same.  A year ago these same five issues came up at
a meeting in Bonn on the development of stove performance benchmarks. A year
later the issues are largely unaddressed, with quite serious consequences
for one of the stoves recently tested, the dissemination of which is
dependant on funding determined by the CO2 and GHG mitigation value. Suppose
these tests are believed?

Failure to address these five fundamental scientific issues, now, at
Aprovecho, can only lead to further complications and disaffection in the
stove developers community.  If they are not addressed, there is a real
danger that their results will be increasingly ignored.

A lot of talented people are willing to contribute to this process.  We must
enhance the store of goodwill that already exists between this eminent group
of peers and deliver superior services to the world.

Sincerely
Crispin