自适应广义数学形态学抑制局部放电窄带干扰  

Adaptive generalized mathematical morphology theory in suppressing narrowband interference of partial discharge

作　　者：曹洁[1] 谢希[1] 吕景顺 

机构地区：[1]兰州理工大学电气工程与信息工程学院,甘肃兰州730050 [2]甘肃省电力科学研究院,甘肃兰州730050

出　　处：《电机与控制学报》2013年第4期33-38,共6页Electric Machines and Control

基　　金：国家自然科学基金(61263031);甘肃省自然科学基金(1010RJZA046)

摘　　要：针对传统数学形态滤波器存在的统计偏倚现象,基于最小均方(LMS)自适应算法调节不同长度的结构元素,提出一种自适应广义数学形态滤波器。在强烈的噪声环境中,采用数学形态学、广义数学形态学和自适应广义数学形态学三种方法对比仿真分析消除连续的周期窄带干扰,并引入信噪比和局部放电脉冲的波形失真均方误差评价指标。仿真分析和现场数据处理结果表明,自适应广义数学形态学的滤波效果优于单一的形态滤波效果,完整地保持了局部放电信号的几何结构特征,信号失真小。分析结果表明,自适应广义数学形态学能有效地抑制连续的周期窄带干扰。Aiming at the existing statistical bias phenomenon of traditional mathematical morphological fil- ter, based on the least mean square （LMS） adaptive algorithm adjusting the structure elements of differ- ent length, an adaptive generalized mathematical morphology filter is proposed. In the strong noise envi- ronment, by comparative analysis of simulation using mathematical morphology, generalized mathematical morphology and adaptive generalized mathematical morphology methods to eliminate the continuous peri- odic narrowband interference. And introducing to signal noise ratio and partial discharge pulse waveform distortion of the mean square error evaluation index, the results of simulation analysis and field data pro- cessing show that adaptive generalized mathematical morphology filter effect is better than signal morphol- ogy filter effect, which completely keep the geometric structure characteristics of partial discharge signals, and the distortion of signals is small. Analysis results show that the adaptive generalized mathematical morphology effectively inhibits the continuous periodic narrowband interference.

关 键 词：变压器 局部放电 广义数学形态学 自适应 连续的周期窄带干扰 

分 类 号：TM855[电气工程—高电压与绝缘技术]