Indonesia | English
Member Login

News Update

Promoting Clean Cookstoves at Atmajaya ...

YDD was invited to promote I-CSI clean biomass cookstoves at Atma Jaya University’s Earth Day exhibition ...
Clean Cookstove Demonstration in ...

Yayasan Dian Desa (YDD) provides promotional support to I-CSI market aggregators in the form of cookstove ...
Launch of the I-CSI Second Stage ...

On November 23, 2015, the Directorate of Bioenergy, the Directorate of New and Renewable Energy and the ...



Our Partners

From Water Boiling Test to Water Heating Test

2016-11-11 09:57:06

CASE STUDY: From Water Boiling Test to Water Heating Test


Clean cookstove initiatives beginning in the early 1980s recognized the need for improvements in the performance of biomass fuel cooking stoves. Recommended test methods were developed at the University of Eindhoven and Bois de Feu, an NGO in France. A review of their methods is instructive because they had resolved several conceptual issues which were reintroduced as part of a Water Boiling Test (WBT) by a VITA-led group in 1985. The original document produced by VITA outlining proposed testing methods presented three complimentary approaches: the Water Boiling Test (WBT), Kitchen Performance Test (KPT) and Controlled Cooking Test (CCT). All of them contain conceptual errors that resulted in the mis-rating of stoves, particularly at low power.

The available test methods at the time were reviewed by Rani, Kandpal and Mullik[1] who noted the problem with the low power metrics and reported on an experiment that showed the amount of fuel required to simmer a pot was independent of the mass of water in the pot, rendering the validity of several reporting metrics questionable. 

In the meantime, the VITA WBT emerged as a dominant testing method in the laboratory environment, particularly in the Western world. Around the same time, India and China adopted quite different methods which they have continued to use until now.

The core claim for the WBT is that it supports the standardization and replicability of controlled laboratory testing to arguably allow for the differentiation of cookstoves at their optimum performance. The performance of a stove is contingent on four factors: (a) the stove, (b) the fuel and its moisture content, (c) the operating procedure or burn cycle, and (d) the cooking vessel or vessels.

The last three factors are highly dependent on the local context, meaning that whatever one chooses for each variable will strongly affect the relative performance rating of ‘the stove’. Performance rating is therefore an evaluation of the combination of factors, not only of the stove. In most cases, a prescribed cooking cycle was written into the test, resulting in performance ratings that are only valid for that particular three-phase burn cycle consisting of a cold start, hot-restart of the fire, and a 45 or 30 minute simmering period. The test ignored emissions before placing the pot on the stove and suffered from many other shortcomings.

Although the WBT has been advertised as simple to conduct and capable of giving useful information to designers, it is not as simple to conduct as either the Indian or Chinese tests and has a larger inherent variability.[2]

Field tests conducted using the VITA CCT were intended to capture an indication of performance based on the cooking practices of local users and “provide an indication of the performance of stoves during actual use”.[3] From the beginning, VITA struggled with, and recognized the challenge of, striking a balance between:

(a)  the standardization of testing methods to ensure replicability across testing locations; and,

(b)  the performance of actual cooking practices in the laboratory environment to better predict the stove’s performance in local usage.

VITA emphasized the distinction between testing “done for local use only (for stove users and others) and testing where the results are intended to be transmitted to other places”.[4]

The core misunderstanding in the stated intention for the CCT is the idea that the performance of a stove is inherent to all circumstances. Test results are only valid for the context in which they were performed. This has been amply validated by the fact that there is little, if any, correlation between laboratory tests in which ‘cooking is simulated’ and cooking in any place that does not ‘cook like that’. A stove tested and optimized using dry fuel will almost always perform badly when tested using wet fuel; whether that is in the same laboratory, another one, or in the field. 

Since the WBT occupied a central the role in laboratory testing in the West, using a fixed fuel, fixed burn cycle and fixed cooking vessel, the results proved untranslatable to different global contexts and unable to transcend local variations in cooking practices. A stove ‘recommended’ by WBT results might not be accepted when used under different circumstances. 

The CCT was supposed to address the issue of translatability to different cooking contexts. However, if the CCT is able to provide useful information, why was it not used in the laboratory by reproducing relevant cooking cycles there? The answer is not clear. However, there is an apparent belief that performance is ‘inherent’: a ‘better stove’ would always out-perform a ‘worse stove’, with cookstove projects opting to promote the ‘better’ stove on the basis of the WBT lab test. 

Because any test conducted ‘differently’ will give a different performance rating, it is critical that a product selection process conducted in the lab or the field be representative of the local context. 

It is incorrect assume that any characteristics of a stove are inherent in the design. For example, the particulate matter (PM) emissions vary considerably from the same stove if the moisture content is altered or the fueling rate is changed. Having two different laboratories agree that a stove performs well with 5% moisture fuel does not indicate how a stove will perform with 15% or 25% moisture fuel, levels at which stove performance may not be optimized. Testing performed out of context carries little information about performance, or gives misleading guidance. 

Any test reports the performance under that set of test conditions. While a properly documented set of conditions is reproducible, it may not be relevant for the reader’s project area.  Where there is no match, there is no useful information conveyed. 

The WBT has been significantly altered from its original version to the present Water Boiling Test Protocol 4.2.3[5] and contains additional output metrics. As a result, test results from different versions are not directly comparable and may not even have the same reporting metrics. Importantly, the core problems with the low power metrics introduced in the original 1985 VITA test have not been addressed. So far as is known, the current version has never been externally reviewed by independent experts. 

Because of the large inherent variability of the WBT protocol and the clear differences between the prescribed lab conditions and the flexible field conditions, there has been a standing conflict between WBT and CCT results. The desire to ensure repeatability and precision in the lab has never been matched in the the practical field. This has been exacerbated by the impracticality of conducting precise tests in field conditions.[6]

The developers of the WBT suggest that the requirement for the stove to perform at low and high power output provide a solid indication of the stove’s performance during actual cooking.[7] While this is theoretically possible, the fixed burn cycle and the poor choice of reporting metrics has resulted in its continued rejection in many circles. The UNFCCC allows for the CCT to be used but has done so without conducting an external review.

Furthermore, concepts inherent to test which made little difference to performance rating in the past have, because of new inventions, come to the fore as significant issues. The most important is the old ‘power station’ idea that the heat transfer efficiency is a good proxy for the fuel efficiency of a stove. The development of stoves that make charcoal concurrent with cooking has resulted in stoves that consume as much or more fuel than a three stove fire. When rated for their ‘heat transfer efficiency’ (a calculation of the percentage of heat released from the fire that enters the pot) there is a mismatch with ‘fuel consumption’ since the heat contained in the charcoal is not released during burning. The WBT rates the efficiency of the use of energy released, not the efficiency of the use of the energy in all the fuel extracted from the forest. Stoves that are 15-20% efficient from the ‘forest’s point of view’ are being been credited with ‘an efficiency’ that is over 50% because they do not burn all the fuel, it is turned into charcoal.

Whatever happens to that charcoal, or smaller amounts produced by other stoves, is independent of the stove that produced it. How to account for this charcoal requires a programmatic decision apart from the performance rating.

The idea that ‘simmering’, or performance at ‘low power’, has an ‘efficiency’ is incorrect, also noted by Rani et al.  Similarly, the idea that the final volume of water in a pot represents ‘cooked food’ is incorrect, particularly for foods cooked in water like potatoes or corn.

It is recognized by developers and proponents of the WBT that it does not provide a complete view of stove performance while performing real cooking tasks in a particular local context. The stated purpose of the WBT is to “measure how efficiently a stove uses fuel to heat water in cooking pot and the quantity of emissions produced while cooking”.[8] The core problem is that even if the metrics were corrected, the idea that a universal test tells us something useful about how a particular stove will perform in any circumstance is fundamentally flawed. This assumption fails to consider, in the lab, the context of that future use. Only contextual performance comparisons are relevant for product selection of biomass stoves.

Alternative testing protocol to WBT

The Water Heating Test (WHT) method developed through the Indonesian Clean Stove Initiative pilot program asserts that cooking practices are usually complex cycles that are not adequately represented by two measurements at high power and a third at low power. In place of the standardized WBT cooking cycle, the pilot program has adopted a method of creating a contextual burn cycle relevant to the Indonesian local cooking experience. This has been paired with the well-known SeTAR Heterogeneous Testing Protocol (HTP) referenced in the IWA 11:2012. The characterization of the burn cycle is achieved by the triplicate replication of various cooking cycles and measuring the average performance. This is done by local cooks using local fuels and pots in a laboratory setting.

Multiple cooking cycles are combined to create an ‘average burn cycle’ called a Technical Test which reproduces the power levels and its variation. It may include the ‘weighting’ of cooking cycles to represent their relative frequency of use.

Thus the WHT is a testing framework based on the HTP into which any locally relevant cooking cycle or set of cycles can be placed. The HTP testing method is the same from country to country, from lab to lab, but everything about the cooking is local. A lab in another country could easily reproduce a documented burn cycle from a catalogue of behaviors. In this manner, a manufacturer can test a stove product intended for a foreign market. This accomplishes the intended goal of the WBT to permit stoves to be tested anywhere, for any market. All that is required is a documented Technical Test and the relevant pots and fuels.
In sum, there are multiple clear differences between the WHT and the WBT:
-    The WHT reports the fuel consumption necessary to replicate a cooking task or a set of cooking tasks; the WBT reports the energy consumption, mathematically treating ‘partially burned fuel remaining’ as ‘unburned raw fuel.’

-      The WHT reports the performance while conducting locally relevant cooking cycles while still being restrained to laboratory conditions, thus replicating ‘average field testing’ in the lab; the WBT reports the performance using a fixed, three stage burn cycle, fixed pot and fuel type.

-      The WHT calculates the performance using measurements appropriate for the metric being rated; the WBT uses somewhat arbitrary metrics taken from various sources, not all of which are relevant, correctly calculated or conceptually valid.

-      The WHT has a low inherent variability because it minimizes the influence of experimental variation in calculations; the WBT has an inherent variability of about 30%, largely because of using ‘water mass remaining’ as a divisor for multiple outputs.

-      The WHT has been informally, externally reviewed multiple times as and whenever it is edited.

It is the aim of the Indonesian initiative that the a contextually accountable testing process will be achieved through the incorporation the social anthropology of cookstove usage into development of the burn cycles employed at the laboratory. There are multiple documented cycles available. As knowledge on local cooking practices increases, it may transpire that some cycles are relevant to other regions and countries.

Any description of cooking behavior can be used in developing a testing protocol. This does not affect the validity of the measurements made, only the applicability to the target population the burn cycle is supposed to represent. The more locally relevant the information on cooking behavior, fuels, foods and pots, the more likely the lab-predicted behavior of candidate stoves will be reflected in the field.

Thus there are two protocols involved: once for the social science and another for laboratory testing  

Download full pdf here

index >>