Research On Water Quality Evaluation And Prediction Method For Water Bloom And Its System Implementation In Beijing Lakes

With the development of society, symptpms of water pollution may be more serious. In recent years, scholars on home and abroad are extremely concerned about eutrophication and water bloom, and study the mechanism and intelligent methods. Setting up a system of water information collecting and analyzing based on study of the water eutrophication appraisal and water bloom prediction has become a focus of water bloom study.This paper first introduces the integrated nutritional status index evaluation method, on this basis, we built the integrated nutritional status index model that is improved by adding improved genetic algorithim and eclectic exciting power function. Based on the water eutrophication evaluation method research,and the gray correlation evaluation model that is improved by adding weight is built.Combined with the water quality characteristics of urban lakes and rivers in Beijing, both of evaluation method are verified to be more closer to the standards proposed by the Ministry of Construction.This evaluation method system is better to regional water environment comprehensive evaluation and classification, and is able to assist in the analysis of water environment such as water eutrophication status. It provides scientific references for water environment planning, water bloom forecasting and control decision-making.The prediction factors are integrated from many physical and chemical indicators of water and environmental factors by the principal component analysis method, the prediction of chlorophyll maximum value is calculated by non-linear approximation of BP neural network, and combining with the improved GM(1,1) model that support modeling for new data generated by accumulation, the corresponding moment is predicted.Further, the chlorophyll change trend is analyzed in the water bloom processing from generation to outbreak, combining with factors affecting water bloom, we adopt the gray model to chlorophyll sub-time processing and make its initial forecast. The non-linear compensation for residual sequence is implemented by BP neural network, and the gray-BP artificial neural network forecasting method is proposed. This method can simulate the complex relation of water bloom factors, take advantages of gray method and artificial neural network, and avoid loss of information by single model. It provides an effective new intelligent method for medium-term and long-term prediction on water-bloom. In order to further enhance the prediction accuracy, the original GM (1, 1) model is improved by using the genetic algorithm.It explores a new intelligent research methods for water bloom high accuracy prediction.Last, the river-lake hydrological information wireless transmission and water bloom early warning systems based on Visual Basic is built. The system adopts the mobile acquisition and transmission terminals, GPSR wireless transmission technology, real-time database online storage technology, Matlab hybrid programming technology and so on. The system can implement Beijing river-lake hydrological information remote monitor, water evaluation and water-bloom early warning automotive.