[Stoves] WBT (never ending)

Crispin Pemberton-Pigott crispinpigott at gmail.com
Thu Jan 31 14:19:48 CST 2013


Dear Frank

 

Crispin > Whatever is happening, the numbers are required in real time.

Frank> Why? Don’t all we need is energy going in – energy left to = energy used during the time of the test? Nice to have real time measurements to see how the stove performs but not needed for the WBT.

 

The WBT is not a very helpful test. That is starting to penetrate the haze of confusion that surrounds the testing of stoves. A WBT (as presently contrived) is a mix of a task-based test and an engineering metrics test and does neither correctly. 

 

If we were to talk about task-based testing (TBT) we would use a begin-and-end- energy value to get an overall number that represents something well defined and well quantified. If we want engineering metrics like ‘heat transfer efficiency we would not do a TBT. That is simply not a productive approach. Both are useful and both are possible, but it is not possible to do both simultaneously and get meaningful answers. You can an answer, but it is not a useful or  correct one. When applying these mixed WBT’s across different platforms the results are skewed against or in favour of a stove or fuel type resulting in unfair comparisons being made. If they are unfair, they are not useful.

 

Crispin>  I think you need to see the SETAR test method in operation. It is not nearly as difficult as you imagine and I am beginning to have more confidence that we can get the fuel analysis as it burns from the combustion products. 

Frank> So many combustion products in all forms of energy bonds that need to be looked at and assembled along with accurate flow measurement, possible leaks, calibration for gas flow and equipment etc.  

 

It is actually much simpler than that. For a start, don’t worry about having to capture literally everything. Suppose you had exotic constituents of smoke that accounted for 0.2% of the combustion. So what? You still have 99.8% of what you are looking for. 

 

Frank> The method to estimate energy left over that have been used are very inaccurate -agree. But we can use much better equipment (CHN analyzer) that will give us the values far more accurate than needed.  

 

That is a good point. There is an old American maxim which is, “Do not put precision and accuracy where it is not needed.” There is no point weighing condensate from a boiling pot to 0.1 g and measuring the kerosene fuel used to 1g. It takes less than 0.1 g of kerosene to boil 1 g of water. It should be the other way round. Should be 0.1g for the fuel and 1.0 for the boiled water. You get my point? If you start with one element of the whole chain giving you numbers ±25% there is little point reducing something else from 1% to 0.1%.

 

 

Frank> If burning liquid fuel why not just weight the fuel to Start – fuel left over? 

 

Again a good point. That is fine unless the thermal efficiency changes with time (as the stove heats up) or perhaps the burner functions different at different temperatures (like a Panda which is quite variable). If you are testing a stove for certification, it is important to know these things. For example a stove in South Africa was tested and it passed the ‘fuel heating test’ which is to say, the temperature of the kerosene did not rise much after 1 hour. But the test was not well conceived. If the fuel was indeed heating at a rate that would have it over-heat (and evaporate without burning) then after perhaps 2 hours it would be very dangerous. One stove passed the test by increasing the storage volume so that absorbed heat couldn’t raise the (larger mass) fuel temperature within one hour. But it was still really dangerous after 2 hours. 

 

If you only measure end points you miss things like that. In another  case merely change the pot diameter tripled the CO output. It was legal with one pot size and illegal with another. If you specify the pot to be used (only) you miss dangerous situations which have already been shown to be real risks.

 

Crispin> Here is an example: if you have a fuel you know is 50% carbon, and 10% is missing (based on what you expected, based on a change in mass) what does that tell you? Quite a bit. It means char is being produced somewhere inside or on the fuel and what is burning is hydrogen. That’s not hard. 

 

Frank> Much easier if you know the CHN of the starting material and the CHN of the final material and weight of both. Very exact measurements. Moisture at the start is easy. Only problem is moisture at the end and I think that can be dealt with. Then we have only the estimated calculations for bond energy to determine the energy used.

 

I would love to have the real times measurements as accurately as possible. The problem is the moisture in the middle – the end we can work out with patience. But if the boiling portion of a WBT is done using wood that dries out entirely and is half charred by the time boiling arrives, what is the energy applied? How do we know what the energy of the burning char/wood mix is during simmering? What are we really trying to find out? The mass of fuel consumed each time a new copy of the task is performed, or the heat transfer efficiency? Or the CO emitted per MJ at different power levels? You must decide up front what we are trying to measure then devise an experiment that determines it exactly. This is done routinely in academia and industry. There is nothing special about stoves that requires us to use vainly imagined metrics and methods. Just get on and do the job properly.

 

Crispin> The big variable is water vapour and when the moisture left the fuel. 

Frank> ‘When’ the moisture left the fuel? We need this for the WBT? Or just total water vapor during the run?

 

If you want to know what the heat transfer efficiency is during some power level of at a certain stage of a cooking cycle, you need to know what the heat generated is. If you dno’t know whether the fuel has dried out yet, you can’t make the calculation.

 

Crispin> This needs to be measured directly in real time and the emissions summed to see what is hydrogen burning and what is fuel moisture evaporating. That is not nearly as difficult a calculation as FTIR requires.

Frank> Seems this is all good info for researching the stove performance but overkill and introduces much more potential error for the WBT test. IMO

 

The WBT is filled with potential for errors already. It is a complex test involving multiple sections. It is a valid way of getting a task-based measurement like total fuel consumption or time to complete. It is not a good way to get a specific energy consumption or time to boil. As soon as you see ‘specific’ in a WBT metric, you know it is probably not a valid metric from an engineering point of view. Many things ‘reported’ by the WBT 3 and 4 series are based on a ‘per litre’ calculation, but if you look into it, there is no theoretical basis for assuming that the measured item is dependent on the number of litres involved. A good one is simmering. Simmering is not dependent on the mass of water in the pot, only the pot hot gas contact area. Boiling water, on the other hand, is dependent on the mass of water in the pot and the pot area. Changing the pot to a smaller or larger one but leaving the mass of water the same changes the boiling rate. Once this is realised, then the value of the output number is…devalued!

 

>I get the feeling people think this should be a quick test and cheap. Always nice if possible. But this is an important test –very important if a stove producer wants to compare his/her stove to others for sale. And that warrants the most accurate and reproducible test we can do. 

 

The WBT? Heavens no. The reason Prof Lloyd does not use the WBT is because the results are ‘irreproducible’. Now imagine if at one lab you cannot reproduce the results consistently, what happens when to try to get another lab to ‘replicate’ the test? Think about this. In order to have a lab replicate the testing of a first lab, you have to be able to say, “These results are comparable within x% and therefore the result has been replicated.” To do that you have to know the precision and accuracy of the test itself and then work out if the other lab has results consistently within the error band. As the WBT’s do not have a known precision nor accuracy, how can any other lab ever claim to have replicated the results? People have been comparing averages of test results which is a completely improper way of showing that a method works precisely. 

 

It is both an inaccurate and irreproducible test for two sets of reasons: it has not been evaluated to know, say, what the 95% confidence index is and calculating the total precision will quickly show that it does not meet our current needs. That is what Jim Jetter and I were discussing. I said that the WBT was analysed by Penn Taylor and that he said it was about 50%. Jim challenged me about that statement. In fact the errors are individually shown in Penn’s thesis but not calculated so I was wrong. But Penn told me the total was about 50% if calculated together. That was before the days of TLUD char making stoves becoming tested frequently and there are serious additional errors encountered when applying the WBT to them. In addition, rice hull gasifiers, also popular, have an additional set of issues. I am comfortable saying the WBT has a precision of 50% so you can quote me instead of Penn. When it is properly calculated and that calculation is reviewed, we can change our opinions (which is all it is at the moment).

 

>Each test may take a week+ to dry, grind and prepare samples for CHN analyzer etc. That’s typical for this type of complexity. It should all boil down to fuel quality as being the biggest variable during testing or round-robin test programs –not test procedure or test equipment.    

 

I am ok with that for fuel-remaining analysis. It is very valuable and after a while we will get ‘typical’ values for certain stoves or procedures and not have to do it again. Or just check occasionally. Prof Lodoysamba invented a machine he makes in Ulaanbaatar that will tell you the ash content of a fuel sample (burned or not) in a few seconds, like, 5 seconds. If it could do MASCON that would be great! (moisture, ash, sulphur, carbon, oxygen, nitrogen)

 

Regards

Crispin

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.bioenergylists.org/pipermail/stoves_lists.bioenergylists.org/attachments/20130131/01333843/attachment.html>


More information about the Stoves mailing list