多尺度形态滤波算法在GIS局部放电检测中的应用  

Application of Multi-scale Mathematical Morphology Filter Algorithm to Partial Discharge Signal Detection for GIS

作　　者：王辉[1] 吴晓春[2] 姚林朋[1] 黄成军[1] 钱勇[1] 江秀臣[1] 

机构地区：[1]上海交通大学电气工程系,上海200240 [2]上海市电力公司,上海200122

出　　处：《高电压技术》2010年第7期1747-1752,共6页High Voltage Engineering

摘　　要：为了抑制GIS局部放电现场检测中的白噪声和各种窄带周期干扰信号,介绍了一种非线性信号处理技术——多尺度形态滤波(multi-scale morphological filter,MMF),提出了一种新的组合形态滤波算法,并应用于GIS局部放电声电联合检测信号处理中。经仿真局放信号处理并且与小波变换处理的结果相比较,在平均误差、均方根偏差以及最大误差等指标上均优于小波的处理结果。通过现场对局部放电引起的超声信号处理结果表明,该算法可以去除超声信号中的噪声干扰,使超声信号直达波起始时刻清晰,结合超高频信号实现对局部放电源准确定位。In order to decrease the white noise and narrow band periodic interference signal contained in the partial discharge （PD） detection for gas insulated switchgear （GIS）,we presented a new nonlinear signal processing method-multi-scale morphological filter（MMF） and the filter algorithm of morphological assemblages. The mentioned method was then applied in the signal processing of PD using acoustic emission （AE） and ultrahigh frequency （UHF） method for GIS. After processing the simulation PD signals and comparison with wavelet transform,we find the morphological filter has better efficiency in the indexes of average error,rms error and maximum error. Moreover,after the processing of AE signals excitated by PD in field,the algorithm shows an excellent efficiency of removing the noises contained in it and gives clear starting time of the nonstop wave of AE signal,and achieves exact PD localization for GIS using the ultrasonic associated with UHF method.

关 键 词：局部放电 GIS 超声 超高频 声电联合检测 数学形态学 多尺度形态滤波 

分 类 号：TM835[电气工程—高电压与绝缘技术]