Prediction And Visualization Of Air Quality Based On Error Back Propagation Neural Network Model

Air is the basis of survival and development for human beings,and also a most important element that all life activities rely on.Air quality may have a direct impact on human health standards and living conditions.Meanwhile,high quality of air can help the city gain a competitive role in the investment compaign.However,with the high pase of industrialization and urbanization,tremendous pressure is being added to environment and air.Motor vehicles are becoming common and large-scale industrial development zones are expanding rapidly.The growing population and increasing consumption of energy and resoures are making the situation even worse.The ecological balance is being damaged seriously by air pollution,which can bring greater harm to human life and production.So it is extremely important to monitor,analyze,forecast and visualize the data of urban air quality,whcih can help us not only master the real time quality of city air,but also record the aspects that may have influences on it.On this data,humans may have a better knowledge of air change trends under different circumstances,which is theoretically important and valuabe for urban planning,construction management,pollution control,environment development and public utilities.In the paper,it is mainy introduced that the urban air quality prediction model we designed and realized to predict the ambient air quality of Tianjin,which is based on Error Back Propagation Neural Network Theory.This theory takes seasons,temprature,weather,wind force and direction into consideratio,to predict the daily thickness of PM2.5?PM10?CO?NO2?O3-8h and SO2 in Tianjin City.The front day concentration of pollutants is selected as test sample data in this model,and normalization and antinormalization are propoed regarding both the input and output data.After massive and repeated experiments,the structural parameters of the network are ascertained.The results show that the effects of the designed model forcasts the urban ambient air quality quite well.At the same time,the air quality index(AQI)is calculated according to the latest promulgated standards of ambient air quality GB3095-2012.And we use the control variates method to analyze the predicted output and explore the relationship between air quality and the weather conditions on the individual meteorological conditions.Therefore,we analyzed the causes of air quality changes and proposed measures to improve it.Finally,we completed the visualization of the predicted results and presented it graphical ly in users interface.