Weihe River Water Quality Evaluation Based On Bp Neural Network Method

The rivers are the most important component of Freshwater that can be used on the land of the landscape sphere, are the closest distance water body near to dwelling and working place of people, are the most easily abtained, impacted and polluted water body by human beings. At present, water quality monitoring and management, controlling water pollution in the world is mainly aimed to rivers water quality, and China is no exception.In the evaluation of natural water quality, the application of BP neural network and genetic algorithm (GA) in artificial intelligence science has been getting more and more concern. Firstly, the present status of research and development of rivers water quality at home and abroad is stated in the paper. Then the basic principle, algorithm and characteristics of BP neural network and GA is described, and analyzed, and BP neural network and GA are introduced in the evaluation of natural water quality. Water quality evaluation modes of Weihe River, based on BP neural network with the weight value and the threshold value Optimized by GA, are established for water quality comprehensive evaluation. The main works in this thesis are as follows:(1) Water environmental pollution situation and the water resources situation in recent years are discussed. Review of the present status of research and development of rivers water quality at home and abroad is given, and the water quality assessment of Weihe river Shaanxi is taken as the main directions.(2) Comprehensive rivers water quality evaluation with water quality parameters can be realized by the techniques for approximating arbitrary nonlinear function mappings between multidimensional spaces. Therefore BP neural network and GA are introduced in the paper for well evaluating comprehensive water quality grade. The basic knowledge about BP neural network and GA, including basic idea, learning algorithm and characteristics, is stated in the paper. Then a brief illustration of the basic constitution and application mode of the BP neural network and GA are given, and the basic processes of the two technologies are emphatically introduced. This algorithm combined GA with BP neural network, which is called GA-BP algorithm and it realized the combination of GA's global search capability and BP algorithm's local optimize performance.(3) The need for water quality evaluation is given after concept of water quality evaluation is stated. According to the basic processes of rivers water quality and experimental results, the advantages and disadvantages of water quality evaluation methods of single factor assessment, principal components analysis (PCA), BP neural network and GA-BP neural network is compared. The results demonstrate that BP neural network can realize nonlinear mapping and reasonable evaluation results can be obtained with several kinds of water quality parameters. However, the BP algorithm is easy to fall into local extremum while the weigh-value of network is optimized using the method of grads-descending. GA-BP model is employed to evaluate water quality, which is based on BP neural network with the weight value and the threshold value Optimized by GA, and is by using the strong global random hunting function of the GA and strong local search, parallel processing, nonlinear optimization and self-organizing and self-learning ability of BP neural network. Therefore, GA-BP model is employed to evaluate water quality, which is theoretically valid, and will be of important value of reference in its practical application, can reflect the water quality of rivers accurately and objectively from the overall, and has good application prospects.