Transformer partial discharge fault diagnosis based on ring area graph entropy features

SHANG Haikun; WANG Zhiming; ZHANG Ranzhe; WANG Mingxue; YAN Jiaxue; JIANG Yuecheng

doi:10.19783/j.cnki.pspc.260051

English Version Journal Login

您当前的位置：

首页 >

文章列表页 >

Transformer partial discharge fault diagnosis based on ring area graph entropy features

更新时间：2026-07-27

- Transformer partial discharge fault diagnosis based on ring area graph entropy features
- Power System Protection and Control Vol. 54, Issue 12, (2026)
- 作者机构：
- 作者简介：
- 基金信息：
- DOI：10.19783/j.cnki.pspc.260051
  CLC：
- Published：2026
- 稿件说明：
移动端阅览
SHANG Haikun, WANG Zhiming, ZHANG Ranzhe, et al. Transformer partial discharge fault diagnosis based on ring area graph entropy features[J]. Power System Protection and Control, 2026, 54(12).
DOI：

SHANG Haikun, WANG Zhiming, ZHANG Ranzhe, et al. Transformer partial discharge fault diagnosis based on ring area graph entropy features[J]. Power System Protection and Control, 2026, 54(12). DOI： 10.19783/j.cnki.pspc.260051.

摘要

针对传统变压器局部放电(partial discharge

PD)故障诊断方法在特征表达不足和小样本条件下识别性能受限的问题，提出了一种基于环面积图谱熵(ring area graph entropy

RAGE)和双层课程元学习(dual-level curriculum meta-learning

DCML)的变压器PD故障诊断方法。首先，利用改进的完备集合经验模态分解(improved complete ensemble empirical mode decomposition with adaptive noise

ICEEMDAN)对非平稳PD信号进行分解，筛选出与原始信号高度相关的有效固有模态分量(intrinsic mode function

IMF)。随后，将有效IMF的功率谱密度投影至极坐标系，计算其面积加权分布概率得到RAGE，从而定量表征PD信号的不确定性。最后，基于PD特征构建DCML模型，实现小样本条件下变压器PD故障的高效识别。实验结果表明，RAGE特征在多种模型对比中展现出显著优势，能够更为精细描述PD信号特征。

Abstract

To address the limitations of traditional transformer partial discharge (PD) fault diagnosis methods in feature representation and their restricted recognition performance under small-sample conditions

a transformer PD fault diagnosis method based on ring area graph entropy (RAGE) and dual-level curriculum meta-learning (DCML) is proposed. First

the non-stationary PD signals are decomposed using improved complete ensemble empirical mode decomposition with adaptive noise (ICEEMDAN)

and the effective intrinsic mode function (IMF) that is highly correlated with the original signals is selected. Subsequently

the power spectral densities of the effective IMFs are projected onto a polar coordinate system

and the area-weighted distribution probabilities are calculated to obtain RAGE

thereby quantitatively characterizing the uncertainty of PD signals. Finally

a DCML model is constructed based on the extracted PD features to achieve efficient identification of transformer PD faults under small-sample conditions. Experimental results show that the proposed RAGE feature presents significant advantages across multiple comparative models and can describe PD signal features more effectively and comprehensively.

关键词

Keywords

references

DE CASTRO B A, DOS SANTOS V V, LUCAS G B, et al. A comparative analysis applied to the partial discharges identification in dry-type transformers by Hall and acoustic emission sensors[J]. Sensors, 2022, 22(5).

ZHENG J, CHEN Z, WANG Q, et al. GIS partial discharge pattern recognition based on time-frequency features and improved convolutional neural network[J]. Energies, 2022, 15(19).

JIAO S, SHI W, LIU Q. Self-adaptive partial discharge denoising based on variational mode decomposition and wavelet packet transform[C] // Proceedings of the 2017 Chinese Automation Congress (CAC), October 20-22, 2017, Jinan, China: 7301-7306.

DRAGOMIRETSKIY K, ZOSSO D. Variational mode decomposition[J]. IEEE Transactions on Signal Processing, 2014, 62(3): 531-544.

CHEN J, HE G, WANG Y, et al. Adaptive PID control for hydraulic turbine regulation systems based on INGWO and BPNN[J]. Protection and Control of Modern Power Systems, 2024, 9(4): 126-146.

ZHENG Y, RAO C, WANG F, et al. Transformer fault diagnosis method based on improved particle swarm optimization and XGBoost in power system[J]. Processes, 2025, 13(10).

KHERIF O, BENMAHAMED Y, TEGUAR M, et al. Accuracy improvement of power transformer faults diagnostic using KNN classifier with decision tree principle[J]. IEEE Access, 2021, 9: 81693-81701.

BIGDELI M, ABU-SIADA A. Clustering of transformer condition using frequency response analysis based on k-means and GOA[J]. Electric Power Systems Research, 2022, 202.

WANG C, WANG Y, WANG Y, et al. Richly connected spatial-temporal graph neural network for rotating machinery fault diagnosis with multi-sensor information fusion[J]. Mechanical Systems and Signal Processing, 2025, 225.

SHANNON C E. A mathematical theory of communication[J]. The Bell System Technical Journal, 1948, 27: 379-423, 623-656.

QUE X, YU Q. Dual-level curriculum meta-learning for noisy few-shot learning tasks[C] // Proceedings of the Thirty-Eighth AAAI Conference on Artificial Intelligence, February 20-27, 2024, Vancouver, Canada: 14740-14748.

DOMINGO D, KAREEM A B, OKWUOSA C N, et al. Transformer core fault diagnosis via current signal analysis with Pearson correlation feature selection[J]. Electronics, 2024, 13(5).

SHANG H, LI Y, XU J, et al. A novel hybrid approach for partial discharge signal detection based on complete ensemble empirical mode decomposition with adaptive noise and approximate entropy[J]. Entropy, 2020, 22(9).

SHANG H, LO K L, LI F. Partial discharge feature extraction based on ensemble empirical mode decomposition and sample entropy[J]. Entropy, 2017, 19(9).

FISHER R A. The use of multiple measurements in taxonomic problems[J]. Annals of Eugenics, 1936, 7(2): 179-188.

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

基于DTCWT与LLE算法的变压器局部放电特高频信号特征参数提取方法

Transformer fault diagnosis method based on XGBoost enhanced by group distributionally robust optimization

基于混合特征选择和IOMA-CNN的变压器故障诊断

基于特征提取与INGO-SVM的变压器故障诊断方法

Maximum hosting capacity evaluation method for distributed generations under fixed overcurrent protection settings

Related Author

HAN Baoguo

MA Chi

LI Jingpeng

WANG Hongfu

LIU Changdao

GAO Tao

ZOU Yang

FU Kaiyu

Related Institution

College of Electrical Engineering and Automation, Fuzhou University

LI Shilin(1. Shenyang Institute of Technology

Liaoning Technical University

Electric Power Research Institute of Guizhou Power Grid Co., Ltd.

College of Advanced Manufacturing Engineering, Chongqing University of Posts & Telecommunications

AI问答

⁰