Research on Classification Algorithm Optimization of Transformer Vibration Eigenvalues

HUANG Jianye; DU Houxian; LIN Shuang; LIU Bingqian; YANG Yan; MA Guoming

doi:10.13296/j.1001-1609.hva.2025.01.021

您当前的位置：

首页 >

文章列表页 >

Research on Classification Algorithm Optimization of Transformer Vibration Eigenvalues

Research & Analysis | 更新时间：2025-04-26

- Research on Classification Algorithm Optimization of Transformer Vibration Eigenvalues
- High Voltage Apparatus Vol. 61, Issue 1, Pages: 181-188(2025)
- 作者机构：
  
  1.国网福建省电力有限公司电力科学研究院，福州　 350007
  2.华北电力大学新能源电力系统国家重点实验室，北京　 102206
- 作者简介：
- 基金信息：
  
  Science and Technology Foundation of State Grid Corporation of China(5700-202121258A-0-0-00)
- DOI：10.13296/j.1001-1609.hva.2025.01.021
  CLC：
- Received：22 August 2024，
  
  Revised：07 November 2024，
  
  Published：16 January 2025
- 稿件说明：
移动端阅览
HUANG Jianye, DU Houxian, LIN Shuang, et al. Research on Classification Algorithm Optimization of Transformer Vibration Eigenvalues[J]. High voltage apparatus, 2025, 61(1): 181-188.
DOI：

HUANG Jianye, DU Houxian, LIN Shuang, et al. Research on Classification Algorithm Optimization of Transformer Vibration Eigenvalues[J]. High voltage apparatus, 2025, 61(1): 181-188. DOI： 10.13296/j.1001-1609.hva.2025.01.021.

摘要

在基于振动特征值法的电力变压器故障检测领域，目前广泛采用基频占比、总谐波畸变率等诊断指标，然而变压器在正常运行时，由于激励源的非线性特性和机械部件共振的影响，将产生200、300 Hz等倍频振动信号，降低了特征值对正常、故障工况的区分度。为解决上述问题，文中搜集了各电压等级变压器正常、故障工况的56组振动数据，检验了传统特征值的分类效果；搭建了3种电压等级的三相电力变压器有限元仿真模型，分析了油箱表面振动分布特征和倍频分量成因；优化了基频占比、总谐波畸变率特征值的计算式，提升了两特征值的分类效果。检验结果表明，优化后两特征值的分类效果分别提升了47%、28%。

Abstract

In the field of fault detection of power transformer based on the vibration signal eigenvalues

such diagnostic indexes as fundamental frequency ratio and total harmonics distortion are widely used at present. However

in case of normal operation of transformer

the multi-frequency signals such as 200 Hz or 300 Hz will be generated due to nonlinear effect of vibration source and resonance of mechanical parts

which reduces the classification of eigenvalues to the normal and faulty conditions. For solving above problem

56 sets of vibration data of transformers at different voltage levels and at normal and fault conditions are collected

and the classification effect of traditional eigenvalues is tested. Moreover

the finite element simulation model of three-phase power transformers with three voltage levels is set up

the vibration distribution characteristics on the oil tank surface and the causes of multi-frequency components are analyzed. The calculation formula of the fundamental frequency ratio and total harmonics distortion is optimized

the classification effect of the two eigenvalues is improved. The Test results show that the classification effect of the two eigenvalues after optimization is improved by 47% and 28%

respectively.

关键词

Keywords

references

Views

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

State Recognition of Transformer Based on SVM Classification of Vibration Signals

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

WU Liping

WU Qiang

FENG Yao

YAO Zhiyu

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

Status Grid Shanxi Yangquan Power Supply Company

College of Electrical Engineering and Information Technology,Sichuan University

⁰