基于自适应滤波和采样的10 kV断路器机械故障声纹采集系统  

10 kV Circuit Breaker Mechanical Fault Voice Print Acquisition System Based on Adaptive Filtering and Sampling

作　　者：丁顺航 李先允[1] 单光瑞 陈兰杭 DING Shunhang;LI Xianyun;SHAN Guangrui;CHEN Lanhang(School of Electrical Engineering,Nanjing University of Engineering,Nanjing 211167,China;State Grid Jiangsu Electric Power Co.,Ltd.Zhenjiang Power Supply Branch,Zhenjiang 212001,China)

机构地区：[1]南京工程学院电气工程学院,南京211167 [2]国网江苏省电力有限公司镇江供电分公司,镇江212001

出　　处：《自动化与仪表》2025年第2期99-105,共7页Automation & Instrumentation

摘　　要：针对基于声纹特征的10 kV断路器机械故障诊断中,环境噪声容易影响识别准确率并延长识别时间,采集系统功耗高的问题。该文提出了一种基于自适应滤波和采样的10 kV断路器机械故障声纹采集系统。通过采用变步长LMS算法改进传统滤波算法,对采集的声音进行降噪处理,以便于后续的故障诊断。将信噪比和数据到达率作为自适应采样的依据,提出一种启发式自适应采样算法。结果表明,该方法具有较高的鲁棒性和处理效率,显著提升了信噪比,同时有效降低了系统功耗。与过采样系统相比,该启发式算法系统功耗约为其71.5%,满足电网对电气设备故障诊断系统的准确性和低功耗要求。In the mechanical fault diagnosis of 10 k V circuit breakers based on voiceprint features,environmental noise can easily affect the recognition accuracy and prolong the recognition time,and the power consumption of the acquisition system is high.This paper proposes a 10 k V circuit breaker mechanical fault voiceprint acquisition system based on adaptive filtering and sampling.By improving the traditional filtering algorithm with the variable step size LMS algorithm,the collected sound is denoised to facilitate subsequent fault diagnosis.The signal-to-noise ratio and data arrival rate are used as the basis for adaptive sampling,and a heuristic adaptive sampling algorithm is proposed.The results show that this method has high robustness and processing efficiency,significantly improves the signal-to-noise ratio,and effectively reduces system power consumption.Compared with the oversampling system,the power consumption of this heuristic algorithm system is about 71.5% of that of the oversampling system,meeting the requirements of the power grid for the accuracy and low power consumption of the electrical equipment fault diagnosis system.

关 键 词：10 kV断路器 机械故障诊断 变步长LMS算法 自适应采样 低功耗 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置]