[Stoves] The WBT allows for comparison between tests and stoves, says Winrock

[Xavier Brandao]

Dear all,

 

Winrock has recently published, on September 2017, on its website, a
toolkit:

https://www.winrock.org/wp-content/uploads/2017/09/Winrock_Cookstove_final_r
educed.pdf 

The toolkit < can be used by various stakeholders, but is primarily aimed at
bringing USG staff and external project developers and implementers
up-to-speed on significant developments in the cookstove sector in recent
years. >

 

According to the toolkit, p 20, < Standardized laboratory testing protocols
and metrics allow for replicability and comparability among tests and across
stoves >, and < The laboratory-based WBT is the most standardized of the
three, and allows for comparison between stoves. >

 

It is unbelievable that Winrock still publishes this kind of claim, today. I
am baffled. This has been time and time again proven wrong, and time and
time again repeated, on this list and elsewhere that the WBT did not allow
to compare stoves. That it should not be used to select stoves, for
programmatic or project purposes. 

Even the very rare supporters of the WBT (I am not sure who they are
actually) agree about that.

 

I am not even talking about the proofs that the WBT is not even good to
develop stoves.

 

In February 2017, a study by Lombardi and al. summarized the issues: 

< Some of WBT critical issues remain unsolved. In particular, the main
weakness of the WBT concerns its real-life relevance. [.] Criticism about
WBT concerns also the repeatability of the protocol, with a number of
researchers claiming that it would need to be reviewed in terms of accuracy.
[.] As a matter of fact, uncertainties related to temperature reading and
vaporisation in the boiling region lead to high variability between test
replicates.

A lot of debate has been made around formulation of metrics, primarily on
thermal efficiency, which is often interpreted as the most immediate and
distinctive stove performance parameter. Studies from Bailis et al.
highlighted how relying on WBT thermal efficiency outputs, regardless of the
relative importance of high and low power cooking tasks among the target
population, can lead to misleading interpretations. Furthermore, Zhang et
al. and Jetter et al. questioned the scientific meaningfulness of thermal
efficiency at simmering. 

Finally, some unsolved issues concerning statistical significance of data
are worth mentioning. WBT 4.2.3 includes "Statistic Lessons for Performance
Testing". The appendix specifies that the minimum number of test replicates
for each model of stove should be three, [.] Wang et al. investigated this
topic using a simplified version of the WBT 3.0 and demonstrated that more
than 5 replicates are likely to be required to avoid impractically large 95%
confidence intervals and that even more replicates may be required to
demonstrate a statistically significant difference in performance between
two or more stoves. >

 

Yet, Winrock, like the GACC and D-Lab and their handbook, are again proving
themselves to be the gravediggers of the already struggling project
developers and implementers, rather than the much needed support they should
be, on matters of testing.

 

This is totally irresponsible from Winrock, and very concerning.

 

I put Elisa Derby in copy of this email.

 

Elisa, I am very much looking forward to hear your views on that.


Best,


Xavier

 



---
L'absence de virus dans ce courrier électronique a été vérifiée par le logiciel antivirus Avast.
https://www.avast.com/antivirus
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.bioenergylists.org/pipermail/stoves_lists.bioenergylists.org/attachments/20171119/d0486b08/attachment.html>