Use of an integrated suite of sensors to simultaneously monitor fuel consumption, air quality, and adoption provides important insights and validates impact metrics for household stoves

Abstract

The rise in sensor-based monitoring in the cookstove sector has been driven by the need for objective quantitative performance evaluation within context of use, and is especially useful if monitoring activities can be conducted by in-country project staff. This research explores the insights achievable from single and cross-sensor analysis following simultaneous in-home deployment of stove temperature loggers, weight-based fuel use loggers, and indoor PM concentration loggers deployed with remote guidance by researchers. Longitudinal performance metrics of an improved metal biomass stove with a chimney within its context of use were obtained using sensor suites consisting of stove temperature sensors (EXACT), household air pollution sensors (HAPEx), and fuel use sensors (FUEL) deployed in 48 households in the Taplejung and Panchthar districts of eastern Nepal. Households in the Taplejung district, comprised mostly of commercial tea houses, had a median reduction in daily household average PM concentration of 45.7% (n = 17) and a median reduction in logged household fuel use of 24.5%, or 2.17 kg/day (n = 15). Households in the Panchthar district comprised of smaller households had a median reduction in daily household average PM concentration of 64.5% (n = 19) and a median reduction in logged household fuel use of 8.13%, or 0.42 kg/day (n = 23). Cross-sensor analysis included use of household PM concentration to verify cooking event initiation and extraneous rises in PM outside of identified cooking events for potential exclusion. Household fuel use profiles were compared to known cooking events to determine whether a household had consistently interacted with the fuel measurement system as instructed, indicating which data were reliable and those that should be flagged. While both cross-sensor analysis and verification methods were examined as potential ways to obtain more information from the gathered data, further development of automated analytics platforms are needed before they can be used as reporting tools by project staff.