Discrimination On Degree Of Algae Bloom With Electronic Tongue

Massive discharge of pollutant containing nutritive material N and P made the eutrophication of water much more serious than ever. Algae blooms among which cyanobacterial bloom is the most popural type, outbreaked frequently in recent years. It put adversely impact on resident’s life and industrial and agricultural production. Consequently, it is urgent for us to develop an effective method for rapid detection of water quality.An electronic tongue was employed to detect the water quality and to grade the polluted water samples in experiments. Specific contents are as follows.First, chemical experiments according to national standard were taken to get N, P contents of the water samples Then samples were classified into three pollution stages such as light, medium and heavy based on national standard etc. to provide reference for discrimination of pollution levels.Second, electronic tongue sensor arrays were optimized. Double-distilled water preparated in the laboratory were selected, according to blank tests, and laboratory temperature controled in 25±2℃. Reproducibility tests were also carried out. and the results showed the detection by electronic tongue has a good reproducibility. Independent tests were determined twice reduplicatly, and the results had good correspondence and the data also had high reliability.Third, the signals and data getting from electronic tongue were optimized via principal component analysis to establish the Fisher discriminatory analysis model. Results showed the identification accuracy is 100% based on the established model. In addition, a BP-ANN model was built to evaluate the different grades of algae bloom, and the identification rate reached 100% with respect to the testing set and training set respectively after the BP net was trained correctly. The results were matched well o chemical experiments.In this study, the response of the sensor arrays and the information fusion technology were applied, in conjunction with sensory discrimination and chemical analysis. Polluted water were distinguished according to comprehensive information of water and analysis of the collected data. The result showed that electronic tongue could provide more intensive and accurate monitoring data and make it feasible to conduct real-time, and online monitoring. It provides a simple and rapid method for monitoring of water quality and establishment of the early warning system of river, lake or other water system. It is of great importance both for human being and environment.