Speaker
Description
Background
Air pollution in office environments resulting from occupancy and activities can affect workers' health and productivity [1,2]. Despite the availability of low-cost air quality and occupancy sensors, there is limited real-time examination of occupant-associated air pollutant exposures in office buildings.
Objectives
This study aims to capture the association between occupancy and activities and occupant exposures to air pollution inside offices.
Methods
The study involved a 2-week field campaign in two Swiss office buildings, one mechanically ventilated and the other having a hybrid ventilation system. Four space types from the two buildings were examined, including open-plan office, singular office, meeting room, and cafeteria. Real-time measurements of air temperature, relative humidity, CO2, size-resolved particles, and TVOC were taken from three stationary locations (two sidewalls and a workstation) in each space. Four office workers were recruited per building to carry a customized vest with CO2/PM monitor and complete activity surveys using smartwatches. By using a cloud-based location platform and staff visual inspection, real-time number of occupants of building and room-level was tracked.
Results
The mean and standard deviation of personal CO2 level were 600±130 ppm, while the background CO2 concentrations of the office were 470±55 ppm. The mean and standard deviation of personal PM2.5 concentrations were 1.3±1.4 µg/m3 while the background PM2.5 levels were 0.8±0.9 µg/m3. The mean and standard deviation of personal PM10 was 11.4±21.9 µg/m3, which was 3-4× higher than background PM10 levels (3.8±4.7 µg/m3). The identified differences between background and personal CO2 and airborne particles indicate the existence of a personal cloud effect associated with human activities, which was more discernible among larger particles.
By applying machine-learning algorithm called decision tree classifier, the study could sort out the most significant attributes for personal exposure and occupancy detection respectively. The study findings showed that the stationary CO2/PM measurement at the desk or wall considered more significant in capturing the personal air pollution exposures than activity-related parameters. For instance, PM level at desk was the best proxy for detecting personal PM exposure. The stationary CO2 monitor at desk/wall was a good proxy for detecting both personal exposures and occupancy. During vigorous activities, consideration of the dynamics of activity was required to better characterize personal PM exposures.
Conclusions
Our preliminary results showed key variables to better define human exposures to indoor air pollution in real office environments that will be further used in developing regression models. This proposition will help building practitioners not only better understand the personal air pollution exposures but also improve the accuracy of contemporary detection on dynamic building occupancy. The study will ultimately contribute to establishing an effective ventilation control while choosing a minimum number of air quality monitors at the best location with the potential for energy savings.
References
[1] Wargocki, P., Wyon, D. P., Baik, Y. K., Clausen, G., & Fanger, P. O. (1999). Perceived air quality, sick building syndrome (SBS) symptoms and productivity in an office with two different pollution loads. Indoor air, 9(3), 165-179.
[2] Chang, T. Y., Graff Zivin, J., Gross, T., & Neidell, M. (2019). The effect of pollution on worker productivity: evidence from call center workers in China. American Economic Journal: Applied Economics, 11(1), 151-72.
| Keywords | Low-cost IAQ sensor, Indoor air quality, Building occupancy, Human exposure, Spatio-temporal variation |
|---|
