Research On Data Analysis Algorithm Of Water Environment Monitoring For The Three Gorges Reservoir Area

The Three Gorges Reservoir Area is not only ecological security barrier in the middle and lower reaches of Yangtze River, but also important strategic reserve base of fresh water resource in our country. After the completion of the 2nd phase water storage, the flow rate and self-purification ability of water got down. So, the Water Environment Monitoring and its data analysis has been a hot research topic which is more and more important. With the support of CSTC's natural science fund, our project team studied the Water Environment Monitoring of the Three Gorges Reservoir Area by WSN since 2008 and a series of research results have been achieved. The main content of this paper is to get a rapid and effective assessment method for water quality analysis.Water environment monitoring data collected by WSN has several characteristics such as large amount, high dimensionality and multiple attribute, et al. It's hard to analyze these data and get the current water quality of a certain area in a short period of time. Therefore, it needs to adopt a fast and effective data processing method to determine water quality. Considering some characteristics of water environment monitoring sample data and the actual demand of the project, analysis algorithms for monitoring data which are suitable for the Three Gorges Reservoir Area have been proposed. Main works are as follows:(1) In order to obtain water environment monitoring data of the reservoir in real time, water environment monitoring platform for the Three Gorges Reservoir Area has been built based on WSN. Relying on this platform, we can complete the acquisition work for part of the monitoring indicators which will be used for data analysis.(2) In order to realize the analysis for monitoring data, characteristics of samples have been studied. Then use MySQL and Excel to clean samples and complete principal component analysis by SPSS. Finally, three water quality indicators selected as characteristic factors which are dissolved oxygen(DO), potassium permanganate index(CODMn) and ammonia(NH3-N)..(3) The method which is most widely used for water quality analysis is fuzzy C-average clustering algorithm(FCM). FCM algorithm is sensitive to the initial iterative center. In order to solve this problem, fast fuzzy C-average clustering(FFCM) algorithm proposed by hard C-average clustering(HCM) and FCM algorithm. Clustering results can be obtained more quickly and efficiently by FFCM.(4) Another problem of FCM can't be solved is that the clustering number is adopted by artificial rule.In order to get the clustering number self-adaptively, a new algorithm called Canopy-FCM proposed with Canopy algorithm. Simulation results show that it can obtain clustering analysis results quickly and effectively by Canopy-FCM algorithm than FCM and FFCM. We can get the clustering number self-adaptively and the initial iterative center more optimized.