TY - BOOK
T1 - Development and Deployment of Low-Cost Sensors for Air Pollution Monitoring
T2 - With a Focus on Mobile and Personal Exposure Applications
AU - Russell, Hugo Savill
PY - 2023/3
Y1 - 2023/3
N2 - Poor air quality is a severe issue affecting global health. In order to combat this, individual exposure must be assessed and exposure hot-spots identified. This would allow links to be drawn between diseases and exposure to specific pollutants, and targeted changes to be made in order to reduce key exposures. Current monitoring methods rely on large, expensive, and sparsely placed outdoor monitoring stations, that are not able to accurately assess air pollution exposure, due to their low coverage. Low-cost sensors (LCS) for air quality monitoring have the potential to solve this issue as they are inexpensive, small and portable real-time monitors, that can be deployed in great numbers and measure in a broad range of environments, including indoors and when mobile. They can serve to both supplement and provide valuable validation basis for exposure models. The projects in this thesis aim to further the field of LCS research, through testing, calibration and novel deployments. Firstly, five models of low-cost particulate matter (PM) sensor were tested in a lab environment to determine their lower limit of detection, response time, and ability to measure transient pollution events, as well the impact of pollutant source, humidity, and temperature. For two out of five models a variable time delay was observed. Secondly, a new calibration method for low-cost gas sensors was developed (Enhanced Ambient Sensing Environment - EASE method) and compared with the primary existing methods, which are field and laboratory calibration. The EASE method was shown to perform better than the laboratory method, whilst requiring less resources, and similar to the field method, but required a fraction of the time. Thirdly, LCS were tested for mobile-monitoring, with walking and vehicle-based studies. The sensor performance did not appear impacted by movement speed and therefore the sensors appear viable for mobile-monitoring. Although, the optical PM sensors perform better than electrochemical gas sensors when measuring in highly variable urban environments. Finally, low-cost PM sensors were shown to perform exceptionally well for real-time measurement of PM2.5 in the Copenhagen Metro system, whilst identifying elevated PM concentrations, of ten to twenty times street level, within it. These were related to design choices in the system. As part of the Metro study, micro-environment classification was tested using data from additional sensors and a classification model, this successfully differentiated micro-environments in the Metro. Conditional calibration, whereby the nodes apply specific calibration models to PM measurements depending on their micro-environment, was introduced as a novel concept, and to our knowledge was applied for the first time in this study. It did not significantly improve sensor performance in this case but is expected to have an impact when used for personal exposure monitoring in varied micro-environments. Building on the projects completed as part of this thesis, new personal exposure monitoring studies, with specifically developed nodes and calibration methods, will commence. These studies will aim to uncover new links between pollution exposure and negative health outcomes. Overall this thesis contributes to the improved use of LCS for better characterisation of personal exposure to health-related pollutants, through personal exposure monitoring and mobile-monitoring, as well as testing of commercial sensors and the development of a new calibration process. Our identification of the Copenhagen Metro as a highly polluted environment will also hopefully lead to improvements in the Metro air quality and perhaps measurement of other metro systems or similar areas with LCS.
AB - Poor air quality is a severe issue affecting global health. In order to combat this, individual exposure must be assessed and exposure hot-spots identified. This would allow links to be drawn between diseases and exposure to specific pollutants, and targeted changes to be made in order to reduce key exposures. Current monitoring methods rely on large, expensive, and sparsely placed outdoor monitoring stations, that are not able to accurately assess air pollution exposure, due to their low coverage. Low-cost sensors (LCS) for air quality monitoring have the potential to solve this issue as they are inexpensive, small and portable real-time monitors, that can be deployed in great numbers and measure in a broad range of environments, including indoors and when mobile. They can serve to both supplement and provide valuable validation basis for exposure models. The projects in this thesis aim to further the field of LCS research, through testing, calibration and novel deployments. Firstly, five models of low-cost particulate matter (PM) sensor were tested in a lab environment to determine their lower limit of detection, response time, and ability to measure transient pollution events, as well the impact of pollutant source, humidity, and temperature. For two out of five models a variable time delay was observed. Secondly, a new calibration method for low-cost gas sensors was developed (Enhanced Ambient Sensing Environment - EASE method) and compared with the primary existing methods, which are field and laboratory calibration. The EASE method was shown to perform better than the laboratory method, whilst requiring less resources, and similar to the field method, but required a fraction of the time. Thirdly, LCS were tested for mobile-monitoring, with walking and vehicle-based studies. The sensor performance did not appear impacted by movement speed and therefore the sensors appear viable for mobile-monitoring. Although, the optical PM sensors perform better than electrochemical gas sensors when measuring in highly variable urban environments. Finally, low-cost PM sensors were shown to perform exceptionally well for real-time measurement of PM2.5 in the Copenhagen Metro system, whilst identifying elevated PM concentrations, of ten to twenty times street level, within it. These were related to design choices in the system. As part of the Metro study, micro-environment classification was tested using data from additional sensors and a classification model, this successfully differentiated micro-environments in the Metro. Conditional calibration, whereby the nodes apply specific calibration models to PM measurements depending on their micro-environment, was introduced as a novel concept, and to our knowledge was applied for the first time in this study. It did not significantly improve sensor performance in this case but is expected to have an impact when used for personal exposure monitoring in varied micro-environments. Building on the projects completed as part of this thesis, new personal exposure monitoring studies, with specifically developed nodes and calibration methods, will commence. These studies will aim to uncover new links between pollution exposure and negative health outcomes. Overall this thesis contributes to the improved use of LCS for better characterisation of personal exposure to health-related pollutants, through personal exposure monitoring and mobile-monitoring, as well as testing of commercial sensors and the development of a new calibration process. Our identification of the Copenhagen Metro as a highly polluted environment will also hopefully lead to improvements in the Metro air quality and perhaps measurement of other metro systems or similar areas with LCS.
KW - low-cost sensors
KW - Personal Exposure Monitoring
KW - mobile monitoring
KW - Air Quality
KW - air pollution monitoring
M3 - Ph.D. thesis
BT - Development and Deployment of Low-Cost Sensors for Air Pollution Monitoring
PB - Århus Universitet
ER -