实测过电压/雷电流信号小波去噪方案探究  

Research on wavelet de-noising scheme for field overvoltage /lightning current signals

作　　者：申元 马仪 王磊 龙羿[2] 姚陈果[2] 米彦[2] 吴昊[2] 

机构地区：[1]云南电力试验研究院(集团)有限公司电力研究院,云南昆明650217 [2]重庆大学输配电装备及系统安全与新技术国家重点实验室,重庆400044

出　　处：《电工电能新技术》2014年第10期64-69,共6页Advanced Technology of Electrical Engineering and Energy

摘　　要：现场监测设备所采集到的过电压/雷电流信号中,蕴含有非常丰富的故障类型信息。但是,现场记录到的实际波形不可避免存在大量白噪声干扰,其在各个频段都占据了部分能量。为了保证和提高监控设备对实测信号分类识别的准确性,需进行预处理以消除过电压/雷电流信号中对分类识别系统无用甚至是有干扰影响的噪声部分。本文将小波阈值去噪算法应用于过电压/雷电流信号领域,应用Matlab软件编写基于该方法的算法程序,并对加噪后的模拟信号进行去噪处理,然后通过评估系数反映各种情况下的去噪效果和还原效果,提出了适合110kV及35kV电压等级的过电压/雷电流信号最优去噪处理方案,并结合某地区35kV、110kV变电站及云南110kV输电线路实测波形证明其去噪效果的有效性。Overvoltage /lightning current signals collected by field monitoring equipment contain very rich information for fault types. However,the field waveforms are not all useful information,since the signals in the presence of a large number of unavoidable white noise in various frequency bands,which have occupied part of the energy. In order to guarantee and improve the monitoring equipment on accuracy of classification and recognition of field signals,before the overvoltage /lightning current signal feature extraction,classification and recognition,the signals must be preprocessed,which aims to eliminate the noise part that has no sense on or does harm to the classification and recognition system of field signals. In this paper,currently popular de-noising method that is wavelet threshold de-noising algorithm is applied to the field of overvoltage signals. Author compiles program based on wavelet algorithm using Matlab software,and uses it to deal with simulation waveforms adding the noise signals,and then by the assessing coefficient to reflect the effects of de-noising and reduction in various situations,therefore,the best de-noising scheme of overvoltage /lightning current signals is put forward. By combining the substation records with this de-noising scheme,the effectiveness of the best de-noising scheme for overvoltage /lightning current signals can be proved.

关 键 词：小波阈值去噪算法 高斯白噪声 过电压/雷电流信号 评估系数 

分 类 号：TM864[电气工程—高电压与绝缘技术]