The Way To Predict Quality Of Surface Water Based On Artificial Neural Network

Water resource is an important material in humans' life and production.China is a country lacking of fresh water.Meanwhile,all the existing water resources in China are polluted to some degree and the quality of surface water are closely related to people's production and quality of life.Water environment system is complex and changeable,therefore,it is not clear at present about mechanism in biological,chemical reaction,which it is difficult or has no way to use mathematical equations to express.The traditional deterministic water quality mathematical model usually requires a large amount of hydrological and water quality parameters in the application process,and these parameters are both in large amount and difficult to measure,which limits the applicability and accuracy of the deterministic water quality mathematical model.Therefore,seeking simple,reliable and accurate method of water quality prediction simulation has been the direction of academic unremitting efforts.Artificial neural network is a large-scale nonlinear dynamic system connected by a large number of neurons.BP artificial neural network is one of the most widely used and successful neural networks.Because of the advantages of artificial neural network,it has been applied to image processing,pattern recognition,nonlinear optimization,expert system,speech processing,automatic target recognition,natural language understanding,robot and other fields,and it has made remarkable achievements.Neural network theory has also become a multi-disciplinary emerging,comprehensive cutting-edge disciplines.In recent years,artificial neural network research has been gradually applied to environmental science,some of which has applied ANN research to water quality eutrophication prediction,environmental impact assessment and water quality prediction.The application of this new technique of artificial neural network to water quality prediction is a starting stage,but it has the characteristics that make it have great potential in this field.Based on the data of PH,dissolved oxygen,COD and ammonium nitrogen in the water quality automatic monitoring station of Heilongjiang River in City Heihe from 2005 year to 2013 year,According to the characteristics of different water quality,this essay respectively sets up a 3-layer BP artificial neural network in MATLAB programming language,and simulation neural network model is formed in the MATLAB platform.The improved L-M algorithm is used to predict the water quality of section of City Heihe Heilongjiang River in 2014 year,which is compared with the actual monitoring data,and the better prediction result is obtained.The verification of the algorithm by the UCI standard data set shows that it is feasible and effective to predict the water quality of surface water with BP artificial neural network.The method can make a rapid and accurate prediction of water pollution,and provide important reference for the protection and prevention of water resources.