Prediction And Evalution Of Lake Water Quality Based On Online Monitoring

Water quality prediction and evaluation is essential for water environment planning and prevention,monitoring water quality and collecting concentration data of water pollutant parameters are the basis of water quality prediction.At present,the monitoring of water environment is mainly carried out by means of fixed samples and laboratory analysis,but this kind of monitoring method has a large workload and high monitoring cost.This paper is based on the mobile aeration equipment online monitoring system for real-time monitoring of water quality,compared with the traditional monitoring methods,online monitoring can save cost and improve the monitoring efficiency.Using quartile method and polynomial curve fitting method was carried out on the water quality sensor collected historical data outliers and correction,use the pre-processed data to establish a LM-BP neural network prediction model of COD,ammonia nitrogen,total nitrogen,total phosphorus concentration data of water quality parameters of four indicators,and evaluate the degree of water pollution.This research has a positive effect on promoting the development of on-line water quality monitoring technology and promoting the wide application of BP neural network model in water quality prediction.This paper mainly studies the lake water quality monitored online by mobile aeration equipment from the following four aspects:1.Conduct reliability comparative analysis between the concentration data of water pollutant parameters collected by the mobile aeration equipment automatic monitoring and the data collected by the laboratory at the same time and in the same place,including data qualification rate analysis,data significance test and data correlation analysis,so as to determine whether the data acquired is accurate and reliable.2.Use the quartile method to identify the abnormal data for the first time,and use the polynomial curve fitting method to identify and correct the abnormal values for the second time,which provided the basic conditions for the establishment of the later prediction model.3.Construct the BP neural network water quality prediction model,preliminarily determine the topology of the neural network,divide the data set,iteratively train the training set,and obtain the parameters which make the network performance better as the initial condition of the water quality prediction model simulation,and output the prediction results of the test set.4.Analyze the prediction results of the BP neural network water quality prediction model,compare them with the BP neural network prediction model without data preprocessing and the prediction results using the multiple linear regression model,and evaluate and analyze the water pollution degree.In this paper,structure of the BP neural network prediction model in water quality prediction showed higher prediction accuracy,with the mobile aeration equipment of the early stage of the online data acquisition,data preprocessing,eliminating outliers and BP neural network is directly related to the selection of network structure and parameters,so it can be used in the prediction research of lake water quality.