声纹识别与支柱瓷绝缘子的振动声学检测  

Vibration and Aconstic Detection for Voiceprint Recognition and Strut Porcelain Insulator

作　　者：钟力强[1] 钟飞[1] 马庆增[1] 黄丰[1] 

机构地区：[1]广东电网公司电力科学研究院,广东广州510080

出　　处：《广东电力》2013年第12期97-101,共5页Guangdong Electric Power

摘　　要：针对目前各种瓷支柱绝缘子在线检测方法存在的不足,提出采用声纹识别算法进行绝缘子检测结果分析。介绍了振动声学法检测法和声纹识别算法的原理,通过声纹分析技术建立声纹特征矢量库,采用该声纹特征矢量库对某变电站支柱瓷绝缘子的振动声学检测结果进行处理。结果显示,正常绝缘子和有缺陷绝缘子的检测结果在声纹上有着不同的特征,声纹分析技术能有效分辨出有缺陷的绝缘子,达到与人工判断同样的效果。Aiming at shortages of present various online detecting method for strut porcelain insulator, this paper proposes to use voiceprint recognition algorithm to analyze insulator detection result. It introduces principles of noise ＆ vibration detec- tion and voiceprint recognition algorithm. On the basis of voiceprint analysis technology, it introduces establishment of voiceprint characteristic vector library which is used for handling noise ＆ vibration detection result of the strut porcelain insu- lator in some substation. The result shows that the normal insulator and defective insulator are respectively provided with dif- ferent characteristics in voiceprint. It also verifies voiceprint analysis technology is able to effectively distinguish defective in- sulator which is of the same effectiveness as artificial Judgement.

关 键 词：振动声学 支柱瓷绝缘子 声纹识别 

分 类 号：TM216[一般工业技术—材料科学与工程]