李云婷,严京海,孙峰,张大伟,夏曦,芮晓光,白鑫鑫,尹文君.基于大数据分析与认知技术的空气质量预报预警平台[J].中国环境管理,2017,9(2):31-36
基于大数据分析与认知技术的空气质量预报预警平台
Air Quality Forecasting Platform Based on Big Data Analytics & Cognitive Technology
  
DOI:10.16868/j.cnki.1674-6252.2017.02.031
中文关键词:  空气质量预报预警  大数据分析  认知技术  案例分析  污染成因分析  决策支持
英文关键词:air quality forecasting and early warning  big data analytics  cognitive technology  case analysis  pollution cause analysis  decision support
基金项目:国家科技支撑计划课题(2014BAC23B03)。
作者单位E-mail
李云婷 北京市环境保护监测中心, 大气颗粒物监测技术北京市重点实验室, 北京 100048  
严京海 北京市环境保护监测中心, 大气颗粒物监测技术北京市重点实验室, 北京 100048  
孙峰 北京市环境保护监测中心, 大气颗粒物监测技术北京市重点实验室, 北京 100048  
张大伟 北京市环境保护监测中心, 大气颗粒物监测技术北京市重点实验室, 北京 100048 zhangdawei@bjmemc.com.cn 
夏曦 IBM中国研究院, 北京 100193  
芮晓光 IBM中国研究院, 北京 100193  
白鑫鑫 IBM中国研究院, 北京 100193  
尹文君 IBM中国研究院, 北京 100193  
摘要点击次数: 934
全文下载次数: 2109
中文摘要:
      为深入认识区域大气污染现象规律,完善并提高城市空气质量预报预警能力,提高大气污染治理决策支持能力,开展城市污染成因分析与空气质量预报预警研究是十分必要的。本文针对环境大数据时代下的城市空气质量预报,提出了一种基于大数据分析与认知技术的专业先进的大气环境业务应用系统体系。该体系基于底层统一的数据资源中心,融合各类不同类型的空气质量监测、不同预报系统的产品数据以及基础辅助数据,建立数据汇交、共享、质控管理机制,通过上层预报预警、综合分析、案例分析、应急决策支持四大子系统,从多模式集合预报结合专家调优支撑高性能预报会商应用,从大数据融合时空关联分析深度挖掘大气复合污染特征与污染成因,从多维度历史污染过程和天气形势全自动化认知分析支撑重污染过程研判,从业务化仿真情景方案与污染溯源助力专业应急决策。最后,通过在北京市环境保护监测中心的系统实现证明体系的高性能、稳定性和实用性。
英文摘要:
      It is necessary to carry out research on urban and regional air pollution causes and air quality forecasting & early warning, in order to better understand the regional air pollution disciplinarian, improve the urban air quality forecasting and early warning ability, and advance the air pollution control decision support capabilities. In this paper, aimed at the urban air quality forecast in the era of environmental big data, a new atmospheric environment service application system based on the big data analytics and cognitive technology is presented. Based on the underlying data resource center, the system integrates all kinds of air quality monitoring data, product data from different forecasting system and basic auxiliary data, establishes the unified mechanism for data exchanging, sharing and quality control. Through the upper four application subsystem – forecasting and early warning, comprehensive analysis, case analysis and emergency decision support, the system supports high performance forecast consultation application from multi-mode ensemble forecasting combined with expert tuning; deeply mines complex air pollution characteristics and causes from spatial correlation analysis on fusion of big data; supports judgments of heavy pollution from automatic cognitive analysis on the multi-dimensional historical pollution process and meteorological trend; assists professional emergency decision making from daily operational simulation on scenarios and pollution source apportion. Finally, the high performance, stability and practicability of the system are proved by system implementation in Beijing MEMC.
HTML  查看全文  查看/发表评论  下载PDF阅读器
关闭