There are significant obstacles to constructing sustainable smart water supply systems in urban cities across the globe. The standard of our drinking water has become an issue of paramount importance in modern society, shifting the focus of municipal planning and policy. Normal physical, chemical, and biological indicators have traditionally been the primary emphasis of urban water quality control. However, because of the delays inherent in using biological indications, serious mishaps, such as wide spread illnesses; have occurred in many major cities. We begin this work by defining the issue at hand and discussing its technical obstacles and open research concerns. We then propose a solution, a risk analysis methodology for the city's water system. Our ability to detect risks and perceive shifts in water quality depends on indicator data gleaned from manufacturing processes. We present an Adaptive Frequency Analysis (Adp- FA) approach to resolve the data by making use of the frequency domain information of indicators for their internal linkages and individual prediction, with the goal of providing results that can be explained. We also look into the scalability of this approach in terms of indicators, locations, and time span. We choose data sets of industrial quality from a Norwegian project in four urban water supply systems (Oslo, Bergen, Strommen, and Aalesund) to use in the application. We put the proposed method through its paces by utilizing it to perform spectrogram, prediction accuracy, and time consumption tests, contrasting it with traditional Artificial Neural Network and Random Forest approaches. The outcomes demonstrate that our approach is superior in most respects. Early warnings of risks associated with water quality in industrial settings and subsequent decision support is possible.

Download