基于太赫兹光谱-统计分析的大气PM_(2.5)监测研究（英文）  

Terahertz-Statistics-Dependent Approach for Monitoring PM_(2.5) in Air

作　　者：姜晨[1] 詹洪磊 李倩[1] 赵昆[1] 张振伟[2] 张存林[2] 

机构地区：[1]中国石油大学(北京)油气光学探测技术北京市重点实验室,北京102249 [2]首都师范大学物理系,北京100048

出　　处：《光谱学与光谱分析》2018年第2期643-649,共7页Spectroscopy and Spectral Analysis

基　　金：the National Key Basic Research Program of China(2014CB744302);the Specially Funded Program on National Key Scientific Instruments and Equipment Development(2012YQ140005);the China Petroleum and Chemical Industry Association Science and Technology Guidance Program(2016-01-07);the National Natural Science Foundation of China(11574401,61405259)

摘　　要：应用太赫兹技术对大气中动力学直径小于2.5μm(PM_(2.5))的细颗粒物进行了定量研究。PM_(2.5)质量和太赫兹吸光度之间存在线性关系,相关系数为0.86。应用主成分分析的方法,可证明随着PM_(2.5)质量的增加,与吸收系数存在相似的趋势。为了提高预测精度,采用偏最小二乘,支持向量机和反向传播人工神经网络对PM_(2.5)进行定量研究。与单一的线性模型相比,统计模型具有较大的预测相关性和较小的误差。对于神经网络模型,训练集与预测集的相关系数和均方根误差分别达到0.999和0.016mg,0.912和0.207mg。因此,THz技术和统计学方法的结合可提供较高精度的预测,作为一种监测PM_(2.5)的有效手段。The study presented a quantitative investigation in monitoring the fine particulate matter with the aerodynamic diameters less than 2.5μm（PM2.5in air with terahertz（THz）technique.Absorption bands were observed and linear relations were obtained between PM2.5 mass and absorbance at selected frequencies with correlation coefficients Rof^0.86.The utilization of principal component analysis（PCA）proved the similar absorption trend in the entire range with increasing PM2.5 mass.In order to improve the prediction precision,we employed other three statistical methods including partial least square（PLS）,support vector machine（SVM）and back propagation artificial neural network（BPANN）for the quantification of PM2.5.Compared with single linear models,statistical models had larger prediction correlation and smaller errors.For BPANN model,R and root-mean square error（RMSE）achieved 0.999 as well as 0.016 mg in training and0.912 as well as 0.207 mg.Therefore,the combination of THz technique and statistical methods represents high precision of prediction and really can act as an effective tool to monitor PM2.5 together with other normal approaches.

关 键 词：太赫兹 吸收 PM2.5 统计方法 

分 类 号：O433[机械工程—光学工程]