Quality Risk Analysis for Sustainable Smart Water Supply Using Data Perception

Di Wu,Hao Wang,Hadi Mohammed,Razak Seidu
DOI: https://doi.org/10.1109/tsusc.2019.2929953
2019-01-01
IEEE Transactions on Sustainable Computing
Abstract:Constructing Sustainable Smart Water Supply systems are facing serious challenges all around the world with the fast expansion of modern cities. Water quality is influencing our life ubiquitously and prioritizing all the urban management. Traditional urban water quality control mostly focused on routine tests of quality indicators, which include physical, chemical, and biological groups. However, the inevitable delay for biological indicators has increased the health risk and leads to accidents such as massive infections in many big cities. In this paper, we first analyze the problem, technical challenges, and research questions. Then, we provide a possible solution by building a risk analysis framework for the urban water supply system. It takes indicator data we collected from industrial processes to perceive water quality changes, and further for risk detection. In order to provide explainable results, we propose an Adaptive Frequency Analysis (Adp-FA) method to resolve the data using indicators’ frequency domain information for their inner relationships and individual prediction. We also investigate the scalability properties of this method from indicator, geography, and time domains. For the application, we select industrial quality data sets collected from a Norwegian project in four different urban water supply systems, as Oslo, Bergen, Strømmen, and Ålesund. We employ the proposed method to test spectrogram, prediction accuracy, and time consumption, comparing with classical Artificial Neural Network and Random Forest methods. The results show our method better perform in most of the aspects. It is feasible to support industrial water quality risk early warnings and further decision support.
What problem does this paper attempt to address?