Enhanced fault detection of an air quality monitoring network

Environmental pollution has adverse consequences on human health and the ecosystem. Among the most dangerous types of pollution is air pollution in urban areas, which has been shown to be strongly linked to higher morbidity and mortality rates. Air pollution can be due to several factors, such as human activities (that produce pollutants such as nitrogen oxides, carbon oxides, and volatile organic compounds), photochemical reactions in the lower atmosphere (that produce ozone), or meteorological conditions that affect the concentrations of dust and particulate matter. The contamination levels of these pollutants need to be maintained below acceptable limits set by the world health organization (WHO) or air quality associations in various areas of the world in order to minimize the impact of these pollutants on humans and the environment. A detection of anomalies in measured air quality data is a crucial step towards improving the monitoring of air quality networks. Therefore, an enhanced fault detection technique of an air quality monitoring network using multiscale principal component analysis (MSPCA)-based on moving window generalized likelihood ratio test (MW-GLRT) is proposed. The presence of measurement noise in the data and model uncertainties degrade the quality of fault detection techniques by increasing the rate of false alarms. Thus, the objective of this paper is to enhance the fault detection of an air quality monitoring network by using wavelet-based multiscale representation of data, which is a powerful feature extraction tool to remove the noises from the data. Multiscale data representation has been used to enhance the fault detection abilities of principal component analysis. The results demonstrate the effectiveness of the MSPCA-based MW-GLRT method over the conventional MSPCA-based GLRT method and both of them provide a good performance compared with the conventional PCA and MSPCA methods.