[Stoves] Test methods for cook stoves

Fri Mar 20 04:36:24 CDT 2015

Dear Frank, Crispin, Philip and everybody,

Thank you for the interesting discussion.

I totally understand your desire to reduce the number of variables for
testing. I would prefer it too if the stoves I design could be tested one
single time in order to have it’s performance assessed, external testing
costs money and each variable you add increases the cost (and time).
The number of variables Frank mentioned (2*3*4*5*6) is not an
overestimation in my view. There is many combinations of fuel and cooking,
but we also have to consider that we are creating a product for the whole
world (well at least 1-3 billion people).

If you design stoves you would probably want to pick one (or a few) markets
for your stove.
Imagine you pick a market, optimize a stove for that market and then have
average external test results for your stove because the ‘global’ test
doesn’t represent the cooking in that market. You are now listed average
out of thousands of products, how will that market ever pick your stove as
most appropriate one? Even though yours is the only one optimized for that
market.

Personally, I love Crispin’s analogy with race tracks so I use it now.
(anyone who missed it, please read to understand the following, I copied it
again underneath my message)

First, I agree with Crispin we need a universal set of equations (or at
least scientific set up) on how certain metrics are calculated.

But besides that imagine if each country/market has it’s own race tracks
(cooking cycles) that explain the power levels, duration and more about how
people cook. I could design and optimize a stove to a specific market and
the external test results would reflect this! Now everybody can comfortably
go and recommend my stove.

Imaging if this set of race tracks was available online as a resource, I
could
- optimize stoves from across the world simply by replicating the race
track in my lab (assuming that I could get the fuel, etc.)
- I can identify all markets that my stove is relevant for, simply by
testing it against various tracks.

This looks to me like an optimal situation, but of course it is not that
easy. How do you pick your race tracks such that they are representative of
the population you describe and not too specific such that the population
you describe is still large enough to represent an interesting market for
manufacturers/distributors to enter.
At this point a lab guy like me has to be quiet and let social scientist,
and field workers decide whether this is actually viable. Although, it
seems from the Indonesia example that it is.

Best regards,
Jiddu

RACE TRACKS
*Like cars, there is a start and warm-up cycle which has to be included in*

*the test because that is part of real life. Power stations and
fixedboilers, however, are usually run continuously and testing them
under?constant conditions? is the norm (and is reasonable).Stoves are
almost never run continuously and certainly not at a constantpower. Thus
performance needs to consider not only the starting cycle, butthe running
cycle with its ups and downs. When testing vehicles, this isdone: the car
is put through a driving cycle on a test bed. It does not haveto go on the
road because the road slope and the power requirement can bemodelled and
reproduced in the lab. All Formula 1 cars are set up this way,per race. The
fuel loaded and suspension and wing settings are tuned in thelab to the
track conditions. The engines are tested by computer operatedservos while
bolted to a frame, not even in the car! The engine and gearboxare ?run
around the circuit? for a whole race to see if the gearing iscorrect and
how much fuel is needed.Sticking with that analogy, we can test the stoves
in the lab using knownconditions and requirements of the ?real track?. Each
race track has a namelike Indianapolis which represents it cycle of
distance, speed, elevationchanges, acceleration and braking requirements.
In vehicle testing eachdriving cycle is given a name and number and
represents a different type ofuse.There is no point in comparing the
performance of an F1 car set up for Monzawith a car set up for Monte Carlo.
You can do it, but it is pointless.  Itdoes not predict performance on any
other track and the teams are well awareof this so they don?t do it.You
cannot pick one race track and say every car must be compared using
thattrack. Which track? Where is the ?average track??  The reason is
obvious: acar can be tuned to the test track, say, a dirt oval in
Charlotte, NorthCarolina, and get an optimal result, yet in real life it
will be used as ataxi on tar roads in Jakarta. If you want to test cars
that will be used asa taxi in Jakarta, you should use a Jakarta driving
cycle, and proper tires.The core lessons are:-          A test protocol can
be created that captures the performance nomatter what the track or
conditions.-          Sets of conditions ? cooking cycles or ?burn cycles?
? can bedescribed and named or numbered.These ?conditions? are not part of
the protocol. The protocol is themeasurement and assessment tool. The
conditions are from the list ofavailable, properly characterised cooking
cycles.Comparing performance means comparing stoves doing the same thing ?
this isagreed. However it should also be agreed that having everything race
on adirt oval in North Carolina is not an appropriate way to characterise
theperformance of a Jakarta taxi.Innovation:  A motorcycle may far
outperform any Jakarta car-taxi, and thatis why there are taxi-motorcycles
in Jakarta. If the goal is to use lessfuel, take a motorcycle. If the goal
is to stay dry, taker a car.  Customers(and cooks) have different goals. We
have to measure appropriately and userelevant conditions when stating
comparative performance.*

*Jiddu Broersma*
*Technology and Organization Officer*

T    +91 413 262 34 37

M   +91 894 055 94 30

W   www.praktidesign.com

A    Spirit Sense, Old Auroville Rd,

       605 104, Bommiyarpalayam,

       Tamil Nadu, India

Previous messages:

Dear Crispin, and Philip,
I like what you and Philip are saying and agree. Because the fuel differs
at different locations there will need to be separate testing anyways so
why not at the same time add Tasks that are common to that local. So we
follow my  outlined procedure and replace the fuel with local available
biomass that works for  the stove being tested and replace the water
boiling task with a real task that is frequently used. There will likely be
several tasks and may even involve several cooking utensils so it could
involve many tests especially if they need to be in trilicate. So to make
it manageable we need to keep the number of real tasks to one to two that
represent typical usage expected and utensil(s) used. The results are still
in biomass/task and time/task. Do we need to add taste quality of the end
results? I don?t think so as that would be difficult.  And we need to
establish a well defined end time when the task is completed. That should
be easy with any task because all we need is the time from match to when
the utensil is up to temperture and add on the time found to be average (as
observed in real World) needed to complete the task. We still need to
complete the task so to calculate the biomass used.
 Is that how you both look at it?
I don?t think we can add emissions to the test package when using real food
- but not sure.
Regards
Frank

Frank Shields
franke at cruzio.com

On Mar 19, 2015, at 4:31 PM, Crispin Pemberton-Pigott <
crispinpigott at outlook.com> wrote:
> Dear Frank
>
> You are getting onto the right path but are defeated by an excess of
opportunities.
> ?FRANK> We are in complete agreement in understanding the variables. But
we can?t test all variables and any test we come up with will not be as
good as the actual person using it (like everything in the World that?s
sold). ?
>
> First, you do not have to test with all the variables covered ? it is
simply not necessary. Stoves are made for markets. Markets have
limitations, constraints. You test for places where you want to sell the
product, just like perfume or shoes.
>
> ?Field tests may be the way to go but who gets to have their stove
tested??
>
> You can do a field test in a lab. You can drag the lab to the field. What
will you do there which you cannot do in another house, or town, or lab? We
have to move away from the idea that a lab test cannot and should not
necessarily represent use. Light bulbs are tested in labs, and represent
use. I already mentioned cars and engines.
>
> ?It should be chosen from a ?most likely to succeed? based on preliminary
tests ? like what I have been describing. IMO:
>
> No problem. Just test it doing what it is going to be doing. Anything
else is a waste of money.
>
> Regards
> Crispin
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.bioenergylists.org/pipermail/stoves_lists.bioenergylists.org/attachments/20150320/e94c7814/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: desconocido.png
Type: image/png
Size: 4837 bytes
Desc: not available
URL: <http://lists.bioenergylists.org/pipermail/stoves_lists.bioenergylists.org/attachments/20150320/e94c7814/attachment.png>