Research On Water Quality Prediction Algorithm Based On Erhai Lake

Water environment is a very important part of the global natural ecological environment,which is closely related to people’s lives.With the development of economy,the pollution also increasing.The erhai-lake,which has always been known for its clear water is not immune.In recent years,due to the emission of pollutants from agriculture,tourism,and industrial production around the erhai-lake,the water quality in the erhai-lake has been severely deteriorated.So the management and early warning of water quality in the erhai-lake have become increasingly important.In this paper,the algorithm of water quality forecasting in erhai-lake is mainly analyzed and researched,and forecasting results have also been comprehensively evaluated.The main research contents and innovations of the paper are as follows:Three kinds of algorithms,entropy method,distance similarity analysis and sensitivity coefficient,are used to determine the main factors that affect the water quality of erhai-lake.Experimental results show that the comprehensive influence of several major chemical factors and non-chemical factors cannot be ignored.In view of the relatively stable characteristics of the erhai-lake water environment,an improved Markov prediction model was constructed.Total nitrogen,total phosphorus and dissolved oxygen are taken as the research object.Finally,simulation results illustrate that prediction accuracy of different chemical factor is more than 83.33%,which is improved significantly compared with traditional Markov forecast.Taking into account the complexity of the water environment system itself,a comprehensive water quality RBF+NEW prediction model is proposed.Temperature,transparency,total nitrogen,total phosphorus,and dissolved oxygen are used as research sample sets to predict the state of water quality in the erhai-lake.The simulation results proved that the RBF+NEW model can not only avoid the instability of the BP neural network model,but also improve the prediction accuracy compared with the radial basis network model.Considering the impact of non-chemical factor content changes on the erhai-lake water quality,the exponential smoothing model and ARIMA model are applied to water quality prediction in the erhai-lake.Chlorophyll a is used as the research object.Simulation results showed that the average relative error was 0.0511,and the exponential smoothing model has a better fit compared with ARIMA model.