The Application Of Artificial Neural Networks In The Evaluation And Prediction Of The Quality Of The Water Environment

Water is not only the life resource, but also the strategic for a state survival and development. That problem of lacking water resources, water logging and water pollution has been the important factors to restrict economic and social development in china. Sustainable utilization of water resource is associated with all human society survival and development. Because hydrographic factor is fuzzy, random and sophisticated, so it is difficult to make a correct water quality evaluation, water quality forecasting with the condition methods. And in the course of study comprehensive evaluation and forecasting, on account of each hydrographic factor is not melting, it is difficult to make a exactly conclusion with the traditional decided methods which will lost some useful information.The application of Artificial neural network on water quality evaluation and water quality modeling are studied in this thesis. The content of the thesis is completed by Capital Normal University and Environment Monitor Center of China.The thesis consists of five chapters.In chapter I , For the purpose of having a clear knowledge of water quality evaluation and water quality forecasting , we present several their newest theories and methods, analyze and compare the methods , then give our opinion of the trend of studying water quality evaluation and water quality forecasting.In chapter II, The specific property of Artificial Neural Networks, as well as the significance of its application in water environmental system , was summarized. Then we give some application of the Artificial Neural Networks in water resource quality. Finally, we introduce BP, RBF and Hopfield water quality evaluation model and BP, L-M and RBF water quality forecasting arithmetic in detail , and prepare and analyze theirnature.In chapter III,A improved BP ANN model on water quality evaluation is presented. The model was set up and applied to the water evaluation with data of the water quality from Environment Monitor Center of China. The results show that this model is simple and convenient to calculate and has better practicability.In chapter IV,we construct a water quality forecasting based the Levenberg-Marguardt algorithm of BP Neutral Network and indexes-time series method. DO COD TN were chose as forecasting target parameters . the one month monitor data of DO COD TN was applied to the model. Comparison between the predicted results of network and the determined results indicated that the L-M network could be used to predict water qualitative.