计算机科学

计算机科学 ›› 2022, Vol. 49 ›› Issue (11A): 211200084-5.doi: 10.11896/jsjkx.211200084

• 交叉&应用 • 上一篇    下一篇

基于深度神经网络与联邦学习的污染物浓度预测二次建模

钱栋炜, 崔阳光, 魏同权   

  1. 华东师范大学计算机科学与技术学院 上海 200062
  • 出版日期:2022-11-10 发布日期:2022-11-21
  • 通讯作者: 魏同权(tqwei@cs.ecnu.edu.cn)
  • 作者简介:(469654952@qq.com)
  • 基金资助:
    国家自然科学基金(62272169)

Secondary Modeling of Pollutant Concentration Prediction Based on Deep Neural Networks with Federal Learning

QIAN Dong-wei, CUI Yang-guang, WEI Tong-quan   

  1. College of Computer Science and Technology,East China Normal University,Shanghai 200062,China
  • Online:2022-11-10 Published:2022-11-21
  • About author:QIAN Dong-wei,born in 1998,postgra-duate,is a member of China Computer Federation.His main research interests include federated learning and semi-supervised learning.
    WEI Tong-quan,born in 1973,Ph.D,associate professor.His main research interests include edge computing,cloud computing,and design automation of intelligent systems and cyber physical systems.
  • Supported by:
    National Natural Science Foundation of China(62272169).

PDF (PC)

317

摘要： 进入新世纪,伴随着我国经济的高速发展,我国很多地区空气污染情况相对严重,同时政府对于空气污染情况的关注度与治理力度也越来越高。当前对于我国空气质量影响最大的是O3,SO2,NO2,CO,PM10,PM2.5这6种污染物,因而对这6种污染物浓度进行预测预报,及时作出相应管控调整就成为了保障居民健康、建设美丽中国的迫切需求。目前污染物预测的主流方案是WRF-CMAQ预测系统,该系统基于污染物物化反应与气象模拟两部分构成。但因为当前对于如臭氧在内的污染物的生成机理等研究还有待深入,WRF-CMAQ模型的预测存在较大误差。因此采用了深度神经网络对污染物浓度进行二次建模的方式,来减少预测误差。同时,采用联邦学习方法,对于多个监测站使用联邦学习进行数据训练,提升模型泛化能力。实验结果表明,相比于一次WRF-CMAQ的一次预测结果,深度神经网络的方案在均方误差值上最多缩小到了3.93%。同时,采用联邦学习的方案相比于单个监测站点在广泛测试中最多提升了68.89%的性能。

关键词: 联邦学习, 深度神经网络, 污染物浓度预测, WRF-CMAQ

Abstract: In the new century,along with the rapid development of Chinese economy,air pollution in many areas of China is relatively serious,while the government is paying more and more attention to air pollution,and its efforts to control air pollution are increasing.Currently,six pollutants that have the greatest impact on China’s air quality are O3,SO2,NO2,CO,PM10,PM2.5.Therefore,predicting and forecasting the concentrations of the six pollutants and making corresponding control adjustments in time have become the urgent needs to protect the health of residents and build a beautiful China.At present,the mainstream solution for pollutant prediction is WRF-CMAQ prediction system,which is based on two parts,physical and chemical reaction of pollutants and meteorological simulation.However,due to the current research on the generation mechanism of pollutants such as ozone is still on the way,the prediction of WRF-CMAQ model has large errors.Therefore,this paper adopts a deep neural network for secondary modeling of pollutant concentrations to reduce the prediction error.At the same time,this paper adopts the federal learning method,and uses federal learning for data training for multiple monitoring stations to improve the model generalization ability.Experiment results show that the deep neural network scheme reduces the mean square error value to at most 3.93% compared to the primary prediction results of one WRF-CMAQ.Moreover,the scheme with federal learning improves the perfor-mance by up to 68.89% compared to a single monitoring site in extensive tests.

Key words: Federated learning, Deep neural networks, Pollutant concentration prediction, WRF-CMAQ

中图分类号: 

  • TP183
[1]HU J,CHEN J,YING Q,et al.1 One-Year Simulation of Ozone and Particulate Matter in China 2 Using WRF/CMAQ Modeling System[J].Atmospheric Chemistry & Physics Discussions,2016,16(16):1-31.
[2]HU J,LI X,HUANG L,et al.Ensemble Prediction of Air Quality Using the WRF/CMAQ Model System for Health Effect Studies in China[J].Atmospheric Chemistry and Physics,2017,17(21):13103-13118.
[3]STEYN D,CHAUMERLIAC N.Recent Advances in Modeling of the Atmospheric Boundary Layer and Land Surface in the Coupled WRF-CMAQ Model[M].Berlin:Springer International Publishing,2016.
[4]PLEIM J,GILLIAM R,APPEL W,et al.Application and Evaluation of High-Resolution WRF-CMAQ with Simple Urban Parameterization [M].Berlin:Springer International Publishing,2014.
[5]BNW A,BHG A,CF J,et al.Simulation of Ozone Formation at Different Elevations in Mountainous Area of Hong Kong Using WRF-CMAQ Model[J].Science of The Total Environment,2015,505(118):939-951.
[6]LEE Y,GERRITY D,LEE M,et al.Prediction of Micropollutant Elimination during Ozonation of Municipal Wastewater Effluents:Use of Kinetic and Water Specific Information[J].Environmental Science & Technology,2013,47(11):5872-5881.
[7]MANIAS D M,SHAMI A.Making a Case for Federated Learning in the Internet of Vehicles and Intelligent Transportation Systems[J].IEEE Network,2021,35(3):88-94.
[8]LI T,SAHU A K,TALWALKAR A,et al.Federated Learning:Challenges,Methods,and Future Directions[J].IEEE Signal Processing Magazine,2020,37(3):50-60.
[9]BROWN D S,FLAGG E.Empirical Prediction of Organic Pollutant Sorption in Natural Sediments[J].Journal of Environmental Quality,1981,10(3):382-386.
[10]ZHANG W,LI X,MA H,et al.Federated Learning for Machine-ry Fault Diagnosis withDynamic Validation and Self-supervision[J].Knowledge-Based Systems,2021,213(1):106679.
[11]MING C,YIN Y,MIN X.Prediction of Hourly Air Pollutant Concentrations near Urban Arterials Using Artificial Neural Network Approach[J].Transportation Research Part D,2009,14(1):32-41.
[12]HINTON G,DENG L,YU D,et al.Deep Neural Networks for Acoustic Modeling in Speech Recognition:The Shared Views of Four Research Groups[J].IEEE Signal Processing Magazine,2012,29(6):82-97.
[13]CIRESAN D,MEIER U,SCHMIDHUBER J.Multi-columnDeep Neural Networks for Image Classification[C]//Computer Vision & Pattern Recognition.IEEE,2012.
[14]EPHRAI M.Speech Enhancement Using a Minimum MeanSquare Error Short-time Spectral Amplitude Estimator[J].IEEE TransAcoust Speech Signal Process,1984,32(6):1109-1121.
[15]KONEN J,MCMAHAM H,YU F,et al.Federated Learning:Strategies for Improving Communication Efficiency[J].arXiv:1610.05492,2016.
[16]SAMARAKOON S,BENNIS M,SAAD W,et al.DistributedFederated Learning for Ultra-Reliable Low-Latency Vehicular Communications[J].IEEE Transactions on Communications,2020,68(2):1146-1159.
[17]REHAK D R,DODDS P,LANNOM L.A Model and Infrastructure for Federated Learning Content Repositories[C]//Proceedings of the 1st International Workshop on Interoperability of Web-based Educational Systems.2005:3-12.
[18]THEODORAS B,CHENRD,MELA T,et al.Federated Learning of Predictive Models from Federated Electronic Health Records[J].International Journal of Medical Informatics,2018,112:59-67.
[19]SUBRAMANYA T,RIGGIO R.Centralized and FederatedLearning for Predictive VNFAutoscaling in Multi-domain 5G Networks and Beyond[J].IEEE Transactions on Network and Service Management,2021,18(1):63-78.
[20]HUR S K,HO C H,KIM J,et al.Systematic Bias of WRF-CMAQ PM10Simulations for Seoul,Korea[J].Atmospheric Environment,2021,244:117904.
[21]CHEN D,XIA L,GUO X,et al.Impact of inter-annual meteorological variation from 2001 to 2015 on the contribution of regional transport to PM2.5 in Beijing,China[J].Atmospheric Environment,2021(D22):118545.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
渝ICP备16013121号-4
地址：重庆市渝北区洪湖西路18号    邮编：401121  
电话：023-63500828（编辑部） 023-67039612（会议/专题） 023-67039623（财务/发票）
E-mail：jsjkx12@163.com
本系统由北京玛格泰克科技发展有限公司设计开发