基于EMD的局部放电窄带干扰抑制算法  

Empirical Mode Decomposition Besed DSI Suppression Algorithm in Partial Discharge Detection

作　　者：姚林朋[1] 黄成军[1] 钱勇[1] 

机构地区：[1]上海交通大学电子信息与电气工程学院,上海200240

出　　处：《电力系统及其自动化学报》2007年第5期33-38,共6页Proceedings of the CSU-EPSA

基　　金：中华电力教育基金会基金项目

摘　　要：以经验模态分解(empirical mode decomposition,EMD)为基础,提出了一种基于包络线及有效值的自适应窄带干扰抑制算法,用以从噪声中提取局放信号。EMD可以自适应地将信号分解成若干阶不同频率段的固有模态函数(intrinsic mode functions,IMF),窄带干扰经分解后在振幅上显示出与局放脉冲明显不同的特征,通过包络线及有效值确定的阈值可以有效地区别开来。仿真及实际处理结果表明:与常规的基于小波变换的窄带抑制算法相比,该算法具有较强的自适应性;能有效地抑制局放信号中的窄带干扰。The theory of empirical mode decomposition （EMD） is briefly introduced in the paper. Furthermore, a novel algorithm based on the envelope and virtual value is proposed for suppressing the discrete spectrum interferences （DSI） in partial discharge （PD） detection. Mixed signals can be decomposed by EMD into various intrinsic mode functions （IMF） with different frequency components. In the obtained IMFs,PD pulses and DSI have different characteristics in amplitude,and can be separated effectively with adaptive thresholds determined by envelope and virtual value. Compared with the classical wavelet based method,the proposed method has better adaptive capacity,and it can eliminate the DSI mixed in the partial discharge effectively ,which is proved by both simulation and practical results.

关 键 词：经验模态分解 固有模态函数 局部放电 窄带干扰 

分 类 号：TM835[电气工程—高电压与绝缘技术]