GIS中SO_2气体紫外光谱检测及数据处理  

Ultraviolet Spectrum Detection of SO_2 Gas in GIS and Data Processing

作　　者：陈鹏[1] 王先培[1] 张军[2] 邓杨[1] 肖伟[1] 胡明宇[1] 周玮[2] 

机构地区：[1]武汉大学电子信息学院,湖北武汉430072 [2]中国电力科学研究院,湖北武汉430074

出　　处：《仪表技术与传感器》2015年第12期42-45,共4页Instrument Technique and Sensor

基　　金：国家自然科学基金资助项目(50677047);湖北省科学条件专项(2013BEC010);中国南方电网科技项目(K-GX2011-019);湖北省科技支撑计划项目(2015BCE074)

摘　　要：气体绝缘组合电器(GIS)设备内发生的局部放电会使绝缘气体SF_6分解,从而导致设备安全性下降。针对此问题选取SO_2作为局放检测的特征标识物,依据紫外光谱检测法对SO_2进行定量检测并搭建实验平台,该方法便于实现现场快速检测。为提高设备精度,首先利用奇异值分解原理对光谱数据进行滤波处理,为减小系统误差选取3个特定波长点的吸光度作为SO_2检测依据进行联合检测体积分数。实验结果表明,应用紫外检测SO_2体积分数误差率在5%以内,满足现场检测要求。Partial discharge of Gas Insulated Switchgear（ GIS） brings SF_6 decomposition,and then the safety of the device declines. To address this problem,the SO_2 was selected as the feature label of partial discharge detection,then quantitatively the SO_2 was detected and the experimental platform based on the ultraviolet spectrum detection method was set up. This method was easy to realize rapid on-site test. In order to improve the accuracy of the equipment,firstly,the principle of singular value decomposition（ SVD） was used to filter spectral data,then the absorbencies of three special wavelength were chosen as basis of SO_2 detection to jointly detect the SO_2 volume fraction to avoid error. The experimental results show that the error of SO_2 volume fraction is within 5percent,so it can satisfy the requirements of on-site test.

关 键 词：紫外光谱 二氧化硫 快速检测 信号去噪 奇异值分解 

分 类 号：TN23[电子电信—物理电子学]