Research On And Implementation Of Air Quality Prediction Based On BP Neural Network

Since China entered the reform and opening up,the rapid development of the economy at the expense of the environment at that time has led to a decline in the quality of the atmospheric environment.At present,China is facing a very serious problem of atmospheric pollution.A accurate prediction of air quality has an vital effect for the environmental protection departments to achieve air pollution control.At present,the on-line air pollution monitoring system established by the environmental protection department only only can make the simple statistical of data,and cannot use the accumulated air pollutant concentration data for further in-depth analysis.However,the existing atmospheric pollution forecast method of numerical models have problems such as difficult to implement,difficult to obtain detailed pollutant emission data being,need expensive hardware resources,and has high computational complexity.Therefore,they are not suitable for the municipal environmental protection department.This paper establishes a statistical model prediction method that analyzes the variation law of historical air pollutant concentration data.The statistical model prediction method has the advantages of saving resource and convenient to implementation,and this model is suitable for the actual needs of environmental protection departments to achieve air quality prediction.Air quality forecast can provide an intuitive quantitative reference for the environmental protection department to formulated various control measures of achieve air pollution.Based on the historical air pollution concentration data,this paper establishes a BP neural network model to learn the statistical laws of air pollutant values to forcast the air quality of a period of time in the future.The main work completed in this paper has the following three aspects: 1.Through analysing the purpose of air quality prediction,an overall framework model for air quality prediction based on statistical modelling is presented.The framework model includes three layers: the acquisition of data,the analysis and processing of data,the feedback of result.The data acquisition is used to obtain the relevant original input factors that involved in the prediction model.the most important for the data analysis and processing layer is to select a appropriate prediction algorithm model and select their input factors to achieve air quality prediction.The result feedback layer is used to encapsulate the prediction results,and The result feedback layer is used to encapsulate the prediction results through data visualization technology,and make a better presentation of the prediction results to the users.2.Making a design of the air quality prediction method based on BP neural network.This method includes four stages: collecting the air pollutant concentration data,processing the data,calculating the air quality index and constructing the prediction network.3.Developed an air quality forecasting system and used it to verified the above models and methods of the air quality forecasting.The experimental results show that the air quality prediction method based on BP neural network designed and implemented in this paper,combined with the developed air quality prediction system,can effectively predict the recent changes in air quality and air pollutant concentration.By collecting concentration data of air pollutants and learning the law of changes in air pollutants to achieve air quality prediction.So it can providing quantitative reference for the environmental protection department to achieve air pollution.