Font Size: a A A

Study On Rapid Response Of Regional Air Pollution Concentration Based On CMAQ And Feed Forward Neural Network

Posted on:2020-07-03Degree:MasterType:Thesis
Country:ChinaCandidate:J C ShiFull Text:PDF
GTID:2381330572464336Subject:Engineering Thermal Physics
Abstract/Summary:PDF Full Text Request
The Community Multiscale Air Quality Modeling System?CMAQ?is an important research tool in the field of atmospheric environmental management and scientific research,and has been applied well at home and abroad.However,the advantages of CMAQ are greatly discounted due to its complexity,professionalism and time-consuming calculation in practical use.Using the pollutants simulated by CMAQ experimental emission control matrix as input data and the feed-forward neural network model based on statistical machine learning,the functional relationship between the proportion of pollutant emission control factors and pollutant concentration can be easily obtained,that is,the fast response model.Based on this model,the fast response model of regional pollutant concentration can be established,which can improve the efficiency of regional pollutant concentration prediction and provide math and statistical basis for formulating emission reduction control schemes in accordance with local conditions and time.The results show that the rapid response model established in this paper has strong reliability and can realize the design function of lightweight CMAQ simulation.This model achieves faster computation speed and better model scalability than traditional RSM model.PM2.5.5 and O3 pollution problem in the YRD region in winter of 2014 were analyzed based on this model.And the result shows that for PM2.5,long-distance trans-regional transmission phenomenon is obvious in the winter of YRD region,and the impact of local pollutant emission is limited.Primary particulate matter emission and secondary precursor especially NH3 have significant impacts on PM2.5.5 concentration in YRD region.Emission control in upper wind direction and cross-regional joint prevention control according to time and local conditions are the solutions to PM2.5pollution in winter of YRD region.The influencing factors of O3 pollutants in winter of YRD region are complex,mainly affected by NOx and VOC.Among them,NOx shows a strong titration effect on O3 in winter,while the concentration of O3 is relatively insensitive to VOC emissions.O3 concentration control should be timely and locally adapted,paying attention to scientific decision-making.
Keywords/Search Tags:CMAQ, Feed-forward neural network, Fast response, External validation, Model application
PDF Full Text Request
Related items