首页 | 官方网站   微博 | 高级检索  
     

基于BP神经网络的VOCs实时源解析方法
引用本文:谢伟,徐娇,林致国,颜滴,张英磊.基于BP神经网络的VOCs实时源解析方法[J].环境工程,2022,40(12):231-238.
作者姓名:谢伟  徐娇  林致国  颜滴  张英磊
作者单位:1. 镇江新区生态环境和应急管理局, 江苏 镇江 212000;
基金项目:大气污染物卫星遥感反演技术开发与应用研究(JSYFB-0501-202012-0237)
摘    要:为了提高挥发性有机物(VOCs)源解析的时效性,基于BP神经网络提出了对VOCs监测数据进行实时源解析的PBP模型。利用验证数据证实了PBP模型可获得与PMF模型准确性相当的源解析结果,并且PBP模型具有不受输入数据量限制且计算速度快的优点。将PBP模型应用于长三角某工业园区监测时长为10个月的VOCs组分在线数据并计算得到高时间分辨率的源解析结果,发现监测点位VOCs主要来源为工业源(25%)>燃煤源(20%)>汽油车尾气源(18%)>溶剂使用源(15%)>柴油车船舶尾气混合源(12%)>油品挥发源(6%)>植物源(5%)。此外,结合风速风向数据对各污染源主要贡献方位进行识别,为VOCs的精确管理和控制提供参考,且具有较大的源解析应用潜力。

关 键 词:挥发性有机物(VOCs)    在线监测    BP神经网络    PMF    来源解析
收稿时间:2022-04-18

RESEARCH ON REAL-TIME SOURCE APPORTIONMENT METHOD OF VOCs BASED ON BP NEURAL NETWORK
Affiliation:1. Zhenjiang New Area Ecological Environment and Emergency Management Bureau, Zhenjiang 212000, China;2. Beijing Capital Air Environmental Science & Technology Co., Ltd, Beijing 100176, China
Abstract:In order to improve the efficiency of VOCs source apportionment work, a PBP model based on BP neural network was proposed, which could analyze VOCs monitoring data in real time. The model was validated with the test dataset, and the results showed that it had the same accuracy as PMF model. Besides, it had the advantages of no limitation from the volume of data and fast computing speed. The PBP model was applied to the hourly resolution online VOCs dataset of an industrial zone in the Yangtze River Delta for 10 months, and the hourly resolution source apportionment results were obtained. The main sources of VOCs at the monitoring site were in the sequence of industrial process source (25%)>coal combustion source (20%)>gasoline vehicle exhaust source (18%)>solvent use source (15%)>mixed diesel vehicle and vessel exhaust source (12%)>fuel volatilization source (6%)>biogenic source (5%). Combined with the wind speed and direction data, the main emission directions of each source were identified, which provided a scientific basis for precise manage and control of VOCs sources. Therefore, it had a strong application potential in source apportionment.
Keywords:
点击此处可从《环境工程》浏览原始摘要信息
点击此处可从《环境工程》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司    京ICP备09084417号-23

京公网安备 11010802026262号