AN ASSESSMENT OF LOW-COST SENSOR PERFORMANCE USED IN URBAN TROPICAL ENVIRONMENT

Abstract

Airborne particulate matter with the aerodynamic diameter size ≤ 2.5µm (PM2.5) is measured by Low-cost sensors (LCS) following the principles of optical method with lesser cost. In this study, the performance and accuracy of LCS (AirBeam) are evaluated against filter sampling, beta attenuation and optical method instruments. Comparing filter based PM2.5, LCS underestimated by 19-21% on average that demonstrates linear correlation (R2 = 0.90). A cut-off point is observed at 34 µg/m3; where LCS underestimates and overestimates below and above the point, respectively. In contrast, LCS demonstrates > 40% overestimation comparing against BAM PM2.5 with a moderate linear relationship (R2 = 0.69-0.73). BAM PM2.5 is underestimated by 35.7 ± 12.6% compared to authentic filter because of semi-volatile organic (SVI) and non-SVI losses from BAM that contributes 9% and 26%, respectively. Comparing LCS, E-Sampler was noticed to be lower by > 50% and demonstrates a non-linear correlation (R2 = 0.90). During rain day, the performance of LCS improves near to authentic filter whereas during non-rain days, LCS underestimates significantly by ~30%. The correlation during smoke dominant (SD) day improves by 5% between filter and LCS. Night-time correlation between E-Sampler and LCS is 5% higher because of more aerosol loading at night.