Investigating fine particulate matter transport in a multi-story house using low-cost sensor measurements and different modeling approaches

This work investigates the transport of fine particulate matter (PM2.5) in a multi-story test house using cooking emissions as a point source. The test house was instrumented with 13 PM2.5 monitors, and the particle sources included pan cooking and air frying, as well as ambient PM2.5 penetration during periods of no indoor activity. In the absence of indoor sources, we observed about 10 % of ambient PM2.5 concentrations penetrating indoors with a time lag of 1 h. Similar peak PM2.5 concentrations were observed for pan frying and air frying of the same food ingredients. A cross-correlation analysis showed that it took 2 to 4 min for kitchen peak concentrations to reach other sensors on the first floor and about 8 min to reach the second floor. PM2.5 concentrations were heterogeneous on the first floor, with non-kitchen areas peaking at 45 % ± 9 % of kitchen levels. Second-floor concentrations were more homogeneous, peaking at 18 % ± 2 % of kitchen levels. Using a typical occupancy scenario, the highest estimated personal PM2.5 exposure (44 %) was experienced in the kitchen/dining area, which accounted for 9 % of the time spent at home. We used three modeling approaches to analyze particle transport throughout the house, with increasing input requirements: a multi-box model, an empirical model, and the NIST CONTAM model. All models predicted time integrated PM2.5 concentrations on the 1st and 2nd floors, with R2 between 0.57 and 0.82 and RMSE from 6 µg m-3 to 11 µg m-3.