State Recognition of Transformer Based on SVM Classification of Vibration Signals

WU Liping; WU Qiang; FENG Yao; YAO Zhiyu; WANG Zhong; ZHAO Lihua

doi:10.13296/j.1001-1609.hva.2019.11.034

您当前的位置：

首页 >

文章列表页 >

State Recognition of Transformer Based on SVM Classification of Vibration Signals

更新时间：2025-04-26

- State Recognition of Transformer Based on SVM Classification of Vibration Signals
- High VoltageApparatus Vol. 55, Issue 11, Pages: 232-238(2019)
- 作者机构：
  
  1. 国网山西省电力公司阳泉供电公司
  2. 四川大学电气信息学院
- 作者简介：
- 基金信息：
- DOI：10.13296/j.1001-1609.hva.2019.11.034
  CLC： TM41
- Published：2019
- 稿件说明：
移动端阅览
[1]武立平,吴强,丰遥,要智宇,王仲,赵莉华.基于振动信号SVM分类的变压器状态识别[J].高压电器,2019,55(11):232-238.

WU Liping, WU Qiang, FENG Yao, et al. State Recognition of Transformer Based on SVM Classification of Vibration Signals[J]. High voltageapparatus, 2019, 55(11): 232-238.
[1]武立平,吴强,丰遥,要智宇,王仲,赵莉华.基于振动信号SVM分类的变压器状态识别[J].高压电器,2019,55(11):232-238. DOI： 10.13296/j.1001-1609.hva.2019.11.034.

WU Liping, WU Qiang, FENG Yao, et al. State Recognition of Transformer Based on SVM Classification of Vibration Signals[J]. High voltageapparatus, 2019, 55(11): 232-238. DOI： 10.13296/j.1001-1609.hva.2019.11.034.

摘要

为提高基于振动信号的变压器状态识别准确率

提取振动信号的峰峰值、均值、偏度等时域特征量和主频率幅值占比、基频幅值占比、50 Hz及其奇次倍频幅值占比等频域特征量构成特征向量

运用支持向量机（SVM）分类方法对变压器进行状态识别。采集了变压器正常与故障运行状态下的振动信号

结合实测信号分析了各特征量随变压器状态的变化

最后对比了单个特征量与多特征量结合对变压器正常短路、正常空载、故障短路、故障空载4种状态的识别准确率。结果表明

结合振动信号多个特征量的SVM分类对变压器的状态识别准确率显著高于单个特征量的识别准确率。

Abstract

To improve the accuracy of transformer state recognition based on vibration signals

the time-domain and frequency-domain features of the vibration signals of a transformer are extracted to constitute the feature vector

including peak-to-peak value

average

skewness

proportion of main frequency amplitude

proportion of fundamental frequency amplitude

and frequency amplitude proportion of 50 Hz and its odd-order multiplication. The classification method of support vector machine（SVM）is used to achieve state recognition of the transformer. The vibration signals of the transformer in normal and fault conditions are collected

and the variation of each characteristic quantity of the signals with transformer state is analyzed. In addition

the recognition accuracy of four kinds of transformer states is compared between application of single characteristic quantity and application of multiple characteristic quantities. The results show that the SVM classification of vibration signals improves the accuracy of transformer state recognition.

关键词

Keywords

references

Views

210

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Vibration Sound Power Measurement Method of UHV Transformer Based on Location of Main Vibration Point

Research on Classification Algorithm Optimization of Transformer Vibration Eigenvalues

Transformer Fault Diagnosis Technique Based on Voiceprint and Vibration

Research on Transformer Fault Diagnosis Method Based on Vibration Noise and BP Neural Network

Prediction Model for Dissolved Gas in Transformer Oil Based on Policy Gradient Optimization

Related Author

LU Zhumao

YU Hua

ZHANG Chao

LIU Guoqiang

HUANG Jianye

DU Houxian

LIN Shuang

YANG Yan

Related Institution

Institute of Electrical Engineering Chinese Academy of Sciences

School of Electronic, Electrical and Communication Engineering University of Chinese Academy of Sciences

State Grid Shanxi Electric Power Company Electric Power Research Institute

State Grid Fujian Electric Power Research Institute

State Key Laboratory of Alternate Electrical Power System with Renewable Energy Sources, North China Electric Power University

⁰