10th International Aerosol Conference September 2 - September 7, 2018 America's Center Convention Complex St. Louis, Missouri, USA
Abstract View
Source Apportionment of Particulate Matter Using Low-Cost Particle Sensors with Co-Located Reference Measurements
DAVID HAGAN, Jesse Kroll, Eben Cross, Joshua Apte, Shahzad Gani, Lea Hildebrandt Ruiz, Sahil Bhandari, Gazala Habib, MIT
Abstract Number: 1096 Working Group: Low-Cost and Portable Sensors
Abstract The past several years has seen the emergence of many low-cost commercial devices for measuring particulate matter (PM) for the monitoring of indoor and outdoor air quality. Although the quality of data from these devices has improved as our understanding of their operation has developed, optical-based particle instruments (OPC’s) suffer from sources of error that limit their value for air quality and source apportionment studies (i.e., lack of composition information, lack of measurements of the smallest particles). Often, deployments of low-cost PM sensors are done in conjunction with additional low-cost, gas-phase sensors which are typically treated as separate instruments. In this work, we present methods (based on a variant of a recurrent neural-network algorithm) for using the additional information provided by gas-phase sensors and available composition information (black carbon, PM1 composition via Aerosol Chemical Speciation Monitor) to enhance our understanding of the particle measurements and increase the utility of deploying low-cost sensor networks. We discuss these results in the context of two multi-year, outdoor sensor deployments in very different environments (Boston, Delhi). In addition, we present results on the effectiveness of using factor analysis as a tool to learn more about pollutant sources using low-cost sensor networks.