变电站继电保护设备故障声源定位关键技术  

Key Technologies for Fault Sound Source Localization of Substation Relay Protection Equipment

作　　者：王佑球 陈永安 陈阳东 方智磊 WANG Youqiu;CHEN Yong'an;CHEN Yangdong;FANG Zhilei(State Grid Anhui Electric Power Co.,Ltd.,Qimen County Power Supply Company,Huangshan 245600,China)

机构地区：[1]国网安徽省电力有限公司祁门县供电公司,安徽黄山245600

出　　处：《电声技术》2025年第2期172-174,共3页Audio Engineering

摘　　要：针对变电站继电保护设备故障声源定位的技术需求,提出一种集成多种处理与分析方法的综合定位技术。通过对故障声源信号进行预处理降低噪声,利用梅尔频率倒谱系数(Mel Frequency Cepstral Coefficient,MFCC)和匹配追踪分解法提取特征,并结合隐马尔科夫模型动态建模设备状态演变。采用时频特征分析精细分解信号幅值、频率、时间分布,提升定位的准确性与健壮性。实验验证表明,该技术在复杂环境中具有高效、精准的定位能力,且具有较强的实用价值。Aiming at the technical requirements of fault sound source location of relay protection equipment in substation,a comprehensive location technology integrating various processing and analysis methods is proposed.Noise is reduced by preprocessing the fault sound source signal,features are extracted by Mel Frequency Cepstral Coefficient(MFCC) and matching pursuit decomposition method,and equipment state evolution is dynamically modeled by combining hidden Markov model.Time-frequency characteristic analysis is used to finely decompose the amplitude,frequency and time distribution of signals,so as to improve the accuracy and robustness of positioning.The experimental results show that this technology has high efficiency and accurate positioning ability in complex environment,and has strong practical value.

关 键 词：继电保护设备 故障声源定位 隐马尔科夫模型 

分 类 号：TN912.3[电子电信—通信与信息系统]