Research On Water Quality Prediction Based On Improved Neural Network

At present,the deterioration of water environment in China is becoming more and more serious.As an important part of natural ecological environment,water environment is highly valued for its protection and treatment.Therefore,water quality prediction has become an important content and topic of water environment research.In view of the water quality characteristics of the main stream of Niulan River in Yunnan Province,considering the complex internal system of water environment,the permanganate index(CODMn)is selected as the comprehensive index of the degree of contamination of organic matter in the water environment,and the process of CODMn concentration change is regarded as A one-dimensional function that changes over time.In this paper,BP neural network water quality prediction model,golden section BP algorithm water quality prediction model and wavelet neural network water quality prediction model are established.The three water quality prediction models are used to study the short-term concentration change of CODMn in the Niulanjiang main stream Qixingqiao section,and the predicted value is calculated.Based on the relative error between the actual values,the error acceptance of the three water quality prediction results is analyzed,and a short-term water quality prediction model of CODMn in the Niulanjiang Qixingqiao section with the best prediction accuracy is obtained.The main contents and research results of this paper are as follows:(1)The BP neural network model is established.The results show that the relative error of the prediction results obtained by the model is within the acceptable range,but there is a problem that the convergence speed is slow and the number of hidden layer nodes is difficult to determine.(2)Aiming at the problems of BP neural network model mentioned above,the water quality prediction model of golden section BP algorithm is established.The model determines the number of hidden layer nodes by golden section method.The results show that the golden section BP algorithm water quality prediction model is in the fitting process and prediction results.The accuracy is improved.(3)Both the BP neural network model and the golden section BP algorithm water quality prediction model have the problems of slow convergence,poor network search ability,and easy to fall into local minimum.Aiming at the shortcomings of the above two models,the wavelet neural network prediction model is established.The results show that the wavelet neural network accelerates the convergence speed,improves the search ability,and solves the problem that the BP neural network is easy to fall into the local minimum.At the same time,the prediction results of the wavelet neural network are more ideal and the error is smaller,which is more suitable for the short-term water quality prediction of the CODMn concentration in the Niulanjiang Qixingqiao section.