Detection of specific contamination events in water distribution system using ultraviolet spectra

In recent years, issues surrounding water quality's security have attracted attention. The use of UV/VIS spectrum, for the in-situ detection of water-quality anomalies, has several advantages: it is free of reagent, has a low cost and facilitates rapid analysis. For these reasons, it is one of the more popular research trends for urban distribution water systems. This paper proposes using the UV-VIS spectrum detection method to resolve the problems of existing methods, which cannot distinguish specific contaminants. We made a pretreatment using orthogonal projection to correct the gap between different batches of spectral data. We then used Partial Least Squares Discriminatory Analysis to extract features from the data set. A comparison of the alarm signal with the optimal threshold, revealed outliers from the training set. Finally, we used the Sequential Bayesian Method to update the probability of Contaminate Intrusion Events and to obtain the alarm sequence. Our results show that this new method is effective in detecting specific contamination, which can assist existing methods to confirm contamination types.