An Analytical Framework for Characterizing Pollutant Source and Sink Mechanisms from Time-Resolved Indoor Air Quality Data

SAEED FARHOODI, Insung Kang, Kaveeta Jagota, Nancy Karpen, Zane Z. Elfessi, Israel Rubinstein, Mohammad Heidarinejad, Brent Stephens, Illinois Institute of Technology

     Abstract Number: 122
     Working Group: Advancing Aerosol Science through Data Analysis Tools

Abstract
The growing availability and accuracy of lower-cost consumer- and commercial-grade air quality sensors have greatly increased the spatial and temporal scales at which indoor exposure measurements can be conducted. Increasing amounts of data also necessitates advances in analysis approaches to provide deeper mechanistic insights beyond concentration measurements alone. Therefore, we developed an analytical framework to characterize indoor pollutant source and sink mechanisms from time-resolved indoor air quality data, based on mass balance principles and open-source signal processing approaches. We applied this framework to data collected from an ongoing randomized, single-blind, parallel-group trial of portable air cleaners (PACs) in the homes of US military veterans with chronic obstructive pulmonary disease (COPD) residing in Chicago, Illinois. The field collected dataset includes concurrent time-resolved measurements of indoor particulate matter (PM) concentrations via low-cost monitors and PAC operational patterns via plug load data loggers. The framework includes algorithms for identifying indoor PM peak and background level to automatically detect and characterize the presence and magnitude of PM sources and emission events, estimate PM concentration at background levels in the absence of indoor sources, and quantify decay or loss events under different PAC operation modes (e.g., off, low, medium, and high fan speed settings). Results reveal that while the indoor PM source in the active group was 166% higher than the sham group, the PM2.5 concentrations at background levels were 19-47% lower in the active group compared to the sham group at the same PAC fan speed settings. The PM2.5 loss rates in the active group increased by ~15%, ~70%, and ~130% in the low, medium, and high fan speed settings, respectively, compared to the baseline (air cleaner off). These findings underscore the utility of the proposed framework in revealing the true benefits of PACs, benefits that would otherwise be obscured without this analytical approach.