Research On Water Organic Quality Anomaly Detection Techniques Based On UV Spectra

Detection of water-quality anomaly is an essential functional requirement for early warning of municipal drinking water distribution system. Rapid detection of water-quality anomaly has great significance to speed up emergency response, and safeguard the safety of distribution system for timely identification of accidental and malevolent contamination events, and data support for warning and strategy making. The main research is about the UV-based detection approach for organic pollution. Performance of detection algorithm for water-quality anomaly is testified with a series of experiments and events simulations. Furthermore, detection of organic event types is discussed.Main work and innovation points are as follows:(1) Based on the theory that the majority of organic materials in water have their own specific UV absoption peaks, the research of the feasibility of UV-based organic pollution anomaly detection is undertaken. PCA-based dimension reducing technique and water-quality anomaly detection algorithm using cluster model are studied. In addition, ten kinds of organic solution such as phenol, trichloroacetic acid (TDA), et al. are analyzed with UV detection, together with PCA and cluster analysis, based on the distribution network simulation experiments. The experiment result indicates the clear division among the test solutions by applying UV and cluster algorithm. Additionally, types of pollutants can be identified according to the principle components of UV spectrum.The feasibility of organic anomaly detection through UV monitoring is testified through the simulation experiment.(2) UV-based water-quality anomaly detection methods are studied. Firstly, PCA-based dimension reducing technique and water-quality anomaly detection algorithm using AR model are analyzed with ROC curve evaluating the performance. In addition, probability of detection and false alarm rate of pollution is analyzed in accordance to UV spectral detection. And the effectiveness of this approach is testified in combination with phenol solution with a density of100μg/l.(3) The limitaitons of water quality anomaly are researched and analyzed, focusing on the lowest limit for UV detection. The characteristics of PCA and AR model are analyzed by analyzing the internal relationship between UV absorption wavelength and specific wavelength in combination with ROC curve and value, providing technical support for subsequent UV spectral detection system. In general, a direct UV based water-quality anomaly detection approach is researched and testified. The analysis and researches focus on the feasibility and the limitaitons of the approach. Several organic pollution simulation experiments in water quality monitoring platform has been completed. UV spectrum data accumulated through the experiments, provides technical support for anomaly detection of municipal drinking water supply system.