ARC Newsletter: Implementing ISO 19867
Testing cookstoves for emissions of harmful pollutants and fuel efficiency has been a specialty of ARC for many years. Up until now we used a method that was referred to as the Water Boiling Test. You can come to our library in Oregon to read some of the original texts that were published in the early 80’s by Volunteers in Technical Assistance.
Now we follow the recently released ISO standard 19867, “Clean cookstoves and clean cooking solutions — Harmonized laboratory test protocols — Part 1: Standard test sequence for emissions and performance, safety and durability”. It is available for purchase at https://www.iso.org/standard/
The standard builds upon the work of many non-profit organizations, universities, and national standards organizations to create a method and reporting guidelines so that data obtained by stove testers is high quality, transparent, and verifiable.
One of the challenges in implementing the standard at our lab has been in obtaining repeatable and accurate measurements of PM2.5 (smoke) by weighing the mass of a sample that was deposited on a filter. With the help of two summer interns from the University of Dayton, Maggie Gallegos and Gunaseharan Murugan, we recently completed a study of the performance of our 0.01 mg resolution balance.
The ISO standard calls for the use of a certified calibration weight to help define the uncertainty of the PM2.5 measurement. During each day of testing the calibration weight was weighed five times. The dataset was collected over a period of two months. At ARC we defined the accuracy of the balance as the difference between the average of the set of measurements of the calibration weight and the certified weight of the calibration weight. We defined the precision as the confidence interval of the average of the set of measurements of the calibration weight as constructed using Student’s t distribution at 90% confidence. The results show that during the period of review the balance had an average precision of +- 0.015 mg and an average accuracy of + 0.018 mg, both of which are better then the 0.05 mg acceptance criteria specified in the standard.