基于可闻声信号的配电变压器内部放电故障识别方法研究  

Research on identification method of distribution transformer internal discharge fault based on audible sound signal

作　　者：金潇[1] 舒乃秋[1] 朱元昊 常怡东 刘亚男[1] 李彦彰 

机构地区：[1]武汉大学电气工程学院,湖北武汉430072 [2]武汉科技大学信息科学与工程学院,湖北武汉430081

出　　处：《武汉大学学报（工学版）》2017年第3期415-421,共7页Engineering Journal of Wuhan University

摘　　要：提出了基于可闻声信号的配电变压器放电故障识别方法.首先,采用希尔伯特-黄变换(HHT)对放电声信号进行处理,提取频带能量作为特征量.然后,利用单分类方法支持向量数据描述(SVDD)对特征量进行识别.通过现场实测配电变压器本体声信号和实验室模拟配电变压器内部放电故障对该方法进行了验证,结果表明该方法能够较好地识别配电变压器内部放电故障.The identification method of distribution transformer discharge fault is proposed based on audible sound signal. First of all, the Hilbert Huang transform （HHT） is used to process the acoustic signal and extract frequency band energy as the feature quantity. Then, the support vector data description （SVDD） of single classification method is taken to identify the feature quantity. The acoustic signal of the distribution transformer is measured; and an internal discharge fault simulation experiment of distribution transformer is made to verify the accuracy of proposed method. The analysis results show that the method identifies the internal discharge fault of distribution transformer effectively.

关 键 词：变压器 放电故障 声信号 希尔伯特-黄变换(HHT) 支持向量数据描述(SVDD) 故障识别 

分 类 号：TM41[电气工程—电器]