[Stoves] Updating your WBT and PEMS/LEMS Spreadsheet

Fri Feb 27 09:27:50 CST 2015

Dear Jiddu

You are in the same position as several others, with the unusual difference that you are in a private company. For that reason I will be careful to share things on this public space and not press you for results which belong to the company.

For everyone else, we have to be careful that we provide ‘fair and balanced’ advice so on one has an unfair advantage, and also to try to agree on reporting metrics that are providing useful information. These corrections to the spreadsheet are only for the purpose of providing as much up-to-date understanding as is publicly available. 

Your questions are:

1. You give different changes for the PEMS and WBT sheets. Are the PEMS and WBT giving different performance results for the exact same measurements (ie. stove) now? 

If you enter the same raw data into the PEMS older version or newer versions of the (unique WBT) spreadsheet you will find that the general layout of the “WBT” page is the same. There are a number of versions of this sheet and they have different corrections in them so if you have an ‘old’ version prior to (I think) serial number 2021 there are a number of changes to be included to bring it up to date and if you wish, to remove some of the old invalid metrics (and some new ones).  Please remove references to the ‘efficiency of simmering’. This has been dismissed by reviewers as long as 30 years ago. No one is yet measuring the ‘efficiency of simmering’ because there is no such thing. You either did, or did not, meet the requirements of simmering. Y/N.

Entering the same raw data into WBT 3.1, 4.1.2, 4.2.3 and any of the PEMS WBT pages gives significantly different ratings of performance. They also do not report the same metrics.  You will notice a large different between 4.1.2 (referenced in the IWA) and 4.2.3 (current under the custodianship of the GACC) any time there is a lot of char produced and where there is a moderate fuel moisture of perhaps 12-15%. Similarly there is a large difference in performance between the current WBT 4.2.3 and PEMS 2013 if there was a lot of water evaporated during boiling (tending to be the case with high power stoves) because as pointed out in the previous message, one uses the initial mass of water (which is correct) and the other uses the final mass (which is not).

Because the differences do not apply equally to each stove, the only way to find out what the actual result is, is to make the changes necessary and have a look. It is too complicated to make a conversion spreadsheet.

2. I am currently fitting the 'raw data' and 'logger data' tabs into our own customised sheet that uses a lot of the HTP protocol. I would like to be able to fit errors and confidence to my data, but I am unaware of the accuracy of all parts of the calculations. Mainly the accuracy of the flow, PM, CO and CO2 I would like to figure out. For CO and CO2 I can do a cross check with gas analyser sampling in the same position, which I'm planning to carry out next week. Do you have any advice/thoughts or previous accuracy measurements you can share? 

The biggest problem we will face with using the test approach of the PEMS (hood x volume corrected for temp and pressure) is that it was designed for fuels are burned continuously, and which contain no oxygen, and which produce no char. It is based on an EPA method for stationary sources and gas furnaces don’t produce charcoal. Thus there are assumptions in the calculations, for example the mass of fuel burned, in which it is expected that all the detected carbon represents all the fuel burned.  Wood contains carbon and hydrogen.  The equipment does not detect water vapour from hydrogen combustion so it cannot tell if you are burning wood gas and thus making charcoal, or not. This has been addressed somewhat in the later versions but the root problem remains. The early PEMS numbers are much less reliable.  But the bottom line is unless you have a combination of gas readings and the mass change, you will not be able to work out, even by estimating, what burned.

How that affects the calculated outputs is this: supposed the volume of gas flow is constant. Suppose the concentration of CO2 and CO is low in the beginning of the fire. How do you know if that is a TLUD burning hydrogen-rich woodgas, or is it just a small charcoal fire?  So the Carbon totals are tracked and summed, but without knowing if the mass charge as large, you don’t know if you have a small fire in the chamber, or a large gasifier operating beautifully.

There is a fundamental difference between trying to measure the emissions from a liquid fuel stove and a biomass stove because a biomass fire almost never burns the fuel ‘evenly’. The ‘hood’ method of emissions measurement assumes at the outset that the fuel is burned continuously the same way, which we all known is almost never the case.

There is an alternative EPA method - a carbon balance method – used to determine emissions for vehicles, but that too has a fatal flaw for us. It assumes that if you detect carbon, you have detected fuel. So, if late in the fire, you are burning mostly charcoal, then the C level of the emissions is quite high relative to the ‘average’ for the wood. Then the heat theoretically available is incorrect because there is almost no hydrogen burning. 

Taking the overall average might not be helpful either because that could only provide the ‘correct answer’ if the firepower, or the mass change (one of them) was constant throughout. The water vapour dilution is a major issue because it contains a combustion product that is not measured. Suppose a lot of left-over fuel at the end is totally dry…or not? If you draw some combustion volume charts using various scenarios you will quickly spot the problems.

What works, then, is a filter on the PM (which gives a total mass), but what to do with the CO?  That can be said to have been (within its limits of detection and quantification) been measured correctly. As long as no inference is made from the CO about what heat was available and when, or the total mass burned, the CO measurement has value. It is not used to correct for the energy lost in the calculation of the heat transfer efficiency.  If you are interested in the HTE you should make that correction as the number is available. Do not include WBT low power in such a measurement – you don’t have enough information to be able to determine the heat transfer efficiency to a hot pot.

Lastly, the matter of system efficiency v.s. heat transfer efficiency (proxy, because no one is measuring the actual heat transfer efficiency) has been discussed already. If the remaining char has no value for the next fire, then it is discarded and cannot be considered mathematically ‘unburned fuel’ because it is ‘consumed’. 

Paul raises the point that more metrics are needed on this matter. Fine. No objections there, but the performance of the product under investigation has to be reported correctly with respect to fuel consumption from the supply, because that is how the Gold Standard and CDM and most projects conceive of it. They have been using the heat transfer efficiency proxy which is quite misleading for a lot of stoves. The more char typically produced, the more misleading the rating, and the scale of the error runs to more than 200% of value.

Assuming you were to correct all the formulas and metrics, the PEMS or some other hood-based approach will correctly give real time performance for ethanol and kerosene and LPG stoves because like cars, they do not make charcoal.  Biomass and coal are fundamentally different.

I am very interested to see what you produce as a working sheet. I had a stab at guessing how much moisture was in the emissions at YDD and applied it to the SeTAR SOP 1.57 heat transfer efficiency spreadsheet which is part of the current Indonesian National Standard. It was a surprise. The heat transfer efficiency curve (which is real time) straightened a lot to an almost horizontal line. I didn’t expect it would be that good, or that it was that constant. The SEET Lab measures humidity now and we are in a position to reasonably estimate the effect of dilution and hydrogen combustion. Still working on it…

The evaporation of water from the fuel acts as a dilutant for all emissions and if it is not tracked, you don’t know by how much. The absolute humidity can exceed 150 g/m3 in a chimney (!) some of which is from combustion and some from drying the fuel. But which? And when? As you will understand implicitly by using a scale and the hood, combining the information provides something close to a picture of what is going on in the fire, within the limitations of the equipment.

If we can first get the structure of the experiment correct, we can then go to the topics of accuracy and precision. The team at CAU is very anxious to tackle this during the coming year.  The SeTAR Team is going address the issue of measuring performance during simmering.

Regards

Crispin

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.bioenergylists.org/pipermail/stoves_lists.bioenergylists.org/attachments/20150227/24ddc9ac/attachment.html>