Research On Water Quality Prediction Model Of Based On RBF Neural Network

In recent years,China's continuous development has also been accompanied by increasing pollution,and many water quality problems have appeared in the Three Gorges reservoir area.According to the survey,it is found that more than 80% of the tributaries of the Three Gorges reservoir area have eutrophication due to the input of the eutrophication water quality of the Yangtze River and the input from surrounding human activities,and many water quality problems such as water blooms have occurred.These problems have seriously affected the safety of the water environment in the Three Gorges reservoir area and the lives of local people,and have become a major obstacle in the process of urban construction.In order to solve these problems,this paper analyzes the water quality data collected from the Zhuxia River,a typical tributary of the Three Gorges.These factors affect the water quality through physical and chemical changes in the water.There are a lot of valuable information behind them.How to extract these valuable information from these data,find out the hidden regular information,and To predict the trend of future water quality changes has become the primary prerequisite for comprehensive management of water pollution,and it is an indispensable part of water quality safety management in the reservoir area.This paper studies the water quality data of Zhuyi River and finds that there is a temporal relationship between the data.The water quality data should be analyzed and predicted by time series analysis.This method of analysis does not mandate that the data have other characteristics outside of the time sequence,as long as there is a temporal dependency between the data in the sequence.By analyzing the water quality data of Zhuyi River studied in this paper,it is found that the water quality data studied in this paper is a nonlinear and unstable time series.Therefore,the ARIMA model for nonlinear data in time series analysis is used.However,this model has certain limitations in nonlinear data prediction.To solve this problem,this paper proposes a water quality prediction model based on RBF neural network.The prediction model uses a Gaussian function as the implicit layer basis function,and the center of the basis function is selected by the orthogonal least squares(OLS)center selection method.After the simulation experiment,the model has good prediction accuracy and can predict the future trend of the data.However,the model also has the problem of falling into local optimum and missing the best model.To this end,the particle swarm optimization algorithm in the optimization algorithm is used to optimize the center and width in the RBF neural network.In this paper,the applicability of ARIMA prediction model,RBF neural network prediction model and RBF neural network prediction model improved by particle swarm optimization algorithm in the prediction of nutrient salt concentration in Zhuyi River is studied and compared.Based on the historical data,the three prediction models are simulated.Through the comparison of the average error and the root mean square error,the improved RBF neural network prediction model based on particle swarm optimization algorithm has strong prediction accuracy and good generalization ability.Suitable for prediction of water quality time series data.The model has certain promotion value,which not only provides more powerful data support for water quality safety management decision-making,but also has important theoretical and practical significance for water environment protection in the reservoir area and guarantee the sustainable development of coastal areas.