AAAR 37th Annual Conference October 14 - October 18, 2019 Oregon Convention Center Portland, Oregon, USA
Abstract View
Using Low-cost Sensor Networks to Identify the Influence of Outdoor Air Quality and Indoor Activities on Indoor Air Quality
JIAYU LI, Aliaksei Hauryliuk, Albert Presto, Carnegie Mellon University
Abstract Number: 904 Working Group: Air Quality Sensors: Low-cost != Low Complexity
Abstract Indoor air quality is critical for human health since an average person spend approximately 22 hours, 90% of daily life, inside buildings. Exposure to polluted indoor air may cause fatigue, headache, nasal irritation, and respiratory infections. Indoor air quality is also highly coupled with outdoor air quality, due to the ventilation procedure operated by the heating, ventilation, and air conditioning (HVAC) system. In this study, a network of real-time affordable multipollutant sensors (RAMPs) was deployed at both indoor and outdoor locations in Pittsburgh to investigate the influence of the outdoor air quality and indoor activities on the indoor air quality. The concentrations of ozone and size-resolved particulate matter (PM) concentrations reported by RAMPs will be analyzed. PM can be generated from various indoor and outdoor activities, and the influence of PM on human health has been studied extensively in previous literature. Ozone, as a primary outdoor pollutant generated by transportation and industrial emissions, can be transported into buildings through ventilation. The patterns of indoor air quality changing with outdoor air quality will be demonstrated for polluted and clean days respectively. Furthermore, the contributions of indoor activities (especially occupancy level) to the indoor air quality will be examined. Principle component analysis and a machine learning algorithm will be used to determine and classify primary parameters that influence indoor air quality. When comparing the indoor air quality with outdoor air quality, conventional methods usually focus on one building only. Using low-cost sensors for such comparison can greatly enlarge the dataset by deploying them in multiple buildings, which is beneficial for obtaining more general conclusions, regardless of the differences among buildings.