Study On Air Quality In Changchun Based On Back Propagation Neural Network

With the gradual acceleration of economic development and the continuous improvement of people’s consumption level,air pollution has became a serious challenge that human society have to face.The serious air pollution not only brings the very big influence to the human’s clothing,food,shelter and transportation,but also has caused the certain hindrance to the ecosystem’s normal movement.Therefore,it is necessary to have an in-depth understanding of the influencing factors of air pollution and to establish a scientific,effective,adaptable and accurate air quality prediction system.The forecast results of air quality can not only help people to do the corresponding preventive measures in advance,but also provide reference for the relevant departments to formulate preventive measures.According to the research status of air quality at home and abroad,based on statistical analysis theory and BP(Back Propagation)neural network,the following three aspects of air quality in Changchun are studied in this paper:First,the spatio-temporal change of air quality in Changchun.Based on the air quality index data from 2016 to 2019 collected from all the air quality monitoring stations in Changchun,this paper empirically analyzes the current situation of air quality change in Changchun from the perspectives of space and time.The results show that moderate or severe air pollution still exists in the heating period of autumn and winter in Changchun,although the overall air quality has improved compared with the previous years.Second,analysis of the factors affecting air quality in Changchun.The annual and monthly trends of each pollutant index in Changchun in the past four years were analyzed.At the same time,the effects of temperature,wind speed and wind direction on Air Quality Index(AQI)and Air Pollutant Index(API)were studied.The results show that:(1)It is confirmed that PM10和 PM2.5 is the primary pollutant of air pollution in Changchun.(2)The influence of air temperature on O3 is obvious,but the influence on CO is the least.(3)Compared with the wind direction,the effect of wind speed on pollutant concentration is more significant.Third,the establishment of the Changchun Air Quality Index(AQI)forecast model.Based on the BP(Back Propagation)neural network theory,the forecast model of Air Quality Index(AQI)in Changchun was established by taking the data of meteorological factors and pollutant concentrations as the input layer and the Air Quality Index(AQI)as the output layer.On this basis,the BP(Back Propagation)neural network is extended to the prediction of other pollutants.The modeling results show that BP neural network can effectively predict Air Quality Index(AQI),but the prediction of pollutants needs to be optimized.In addition,the paper puts forward the following three suggestions:first,optimize the energy structure,reduce the emission of waste gas;Second:strengthen the control of Straw burning ban,cultivate and develop new energy;Third:increase the greening efforts,continue to improve air quality.