Acoustical Signal Feature Extraction Method on Mechanical Fault of High Voltage Circuit Breaker Based on Nonparametric Test

YI Lin; XIA Xiaofei; PENG Zaixing; RAO Xiajin; LU Yufeng; SU Yi; LIU Kai

doi:10.13296/j.1001-1609.hva.2021.10.011

您当前的位置：

首页 >

文章列表页 >

Acoustical Signal Feature Extraction Method on Mechanical Fault of High Voltage Circuit Breaker Based on Nonparametric Test

更新时间：2025-04-26

- Acoustical Signal Feature Extraction Method on Mechanical Fault of High Voltage Circuit Breaker Based on Nonparametric Test
- High VoltageApparatus Vol. 57, Issue 10, Pages: 83-90(2021)
- 作者机构：
  
  1. 南方电网科学研究院有限责任公司
  2. 广西电网有限责任公司电力科学研究院
- 作者简介：
- 基金信息：
- DOI：10.13296/j.1001-1609.hva.2021.10.011
  CLC： TM561
- Published：2021
- 稿件说明：
移动端阅览
YI Lin, XIA Xiaofei, PENG Zaixing, et al. Acoustical Signal Feature Extraction Method on Mechanical Fault of High Voltage Circuit Breaker Based on Nonparametric Test[J]. High voltageapparatus, 2021, 57(10): 83-90.
DOI：

YI Lin, XIA Xiaofei, PENG Zaixing, et al. Acoustical Signal Feature Extraction Method on Mechanical Fault of High Voltage Circuit Breaker Based on Nonparametric Test[J]. High voltageapparatus, 2021, 57(10): 83-90. DOI： 10.13296/j.1001-1609.hva.2021.10.011.

摘要

基于非参数检验的声信号时域特征识别是一种高识别率、小时间开销的高压断路器机械故障诊断方法

但也有无效区间长、特征测量值和检验方法有待优化等问题。对于高压断路器机械故障声信号

根据非参数检验的效果剔除低效区段;利用散布矩阵对比特征测量值的分类能力;搭建高压断路器机械故障试验平台

对采集到的声信号分别进行4种非参数检验

选取特征值并进行机器学习。声信号的纯噪音部分和余震部分被剔除

从而得到有效时域区段;在8种特征测量值中

各种熵的分类能力明显高于均值等传统特征测量值

且Renyi熵的分类能力最好;相比于绝对值均值

Renyi熵作为特征测量值时3种非参数检验方法的诊断识别率提高了10%。所得结果表明:Wilcoxon秩和检验选出的Renyi熵特征值有着最高的诊断识别率。

Abstract

Acoustical signal time-domain feature recognition based on nonparametric test is a kind of mechanical fault diagnosis method of high voltage circuit breaker with high recognition rate and small time overheadand

on the other hand

also has such problems as long invalid interval and further optimization of feature measurement and test method. In view of the mechanical fault acoustical signal of high voltage circuit breaker

the low efficiency sections are removed in accordance with the effect of nonparametric test. The classification ability of feature measurements is compared by using the distribution matrix. A mechanical fault test platform for HV circuit breaker is set up and four kinds of nonparametric tests for the collected acoustical signals are performed. Eigenvalues are extracted and the machine learning is made. The pure noise part and aftershock part of the acoustical signals are eliminated so to obtain the effective time domain segment. Among the 8 feature measurements

the classification abilities of various entropies are significantly higher than that of such traditional feature measurements as the mean value and Renyi entropy has the best classification ability.Compared with absolute mean value

the diagnostic recognition rate of the three nonparameter test methods is increased by 10% when the Renyi entropy is used as feature measurement. The obtained results show that the Renyi entropy eigenvalue selected by Wilcoxon rank and the inspection has the highest diagnostic recognition rate.

关键词

Keywords

references

Views

143

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Research on Machinery Fault Diagnosis of High Voltage Circuit Breaker Based on Feature Evaluation

Application of Hybrid Cuckoo Algorithm in High Voltage Circuit Breaker Fault Diagnosis

Mechanical Fault Diagnosis for Spring Mechanism of High-voltage Vacuum Circuit Breaker Based on PSO-LSSVM

Multi-signal Joint Identification and Intelligent Diagnosis of High Voltage Circuit Breaker Faults Based on VMD-BiLSTM

Mechanical Defect Detection Method of High Voltage Circuit Breaker Based on Motion Trajectory Detection and Support Vector Machine

Related Author

HUANG Jian

XU Qidan

ZHANG Xinyan

LI Xiaozhu

ZHAO Liwei

JIANG Hailong

LI Tongqing

CHENG Hao

Related Institution

Beijing Institute of Automatic Control Equipment

College of Electrical Engineering,Xinjiang University

Zhuzhou CRRC Times Electric Co.,Ltd.

Key Laboratory of Special Electric Machines and High Voltage Apparatus in the Ministry of Education,Shenyang University of Technology

Xi’an High Voltage Apparatus Research Institute Co., Ltd.

⁰