基于改进卷积变分自编码器的小样本配电变压器故障诊断方法

高 伟; 姜一涛; 李奕锋

doi:10.19783/j.cnki.pspc.260296

欢迎来到能源电力期刊网！

English Version 期刊登录

您当前的位置：

首页 >

文章列表页 >

基于改进卷积变分自编码器的小样本配电变压器故障诊断方法

应用研究 | 更新时间：2026-09-28

- 基于改进卷积变分自编码器的小样本配电变压器故障诊断方法
- Small-sample distribution transformer fault diagnosis based on improved convolutional variational autoencoder
- 电力系统保护与控制 2026年54卷第19期
- 作者机构：
- 作者简介：
- 基金信息：
- DOI：10.19783/j.cnki.pspc.260296
  中图分类号：
- 纸质出版：2026
- 稿件说明：
移动端阅览
高伟, 姜一涛, 李奕锋. 基于改进卷积变分自编码器的小样本配电变压器故障诊断方法[J]. 电力系统保护与控制, 2026,54(19).

GAO Wei, JIANG Yitao, LI Yifeng. Small-sample distribution transformer fault diagnosis based on improved convolutional variational autoencoder[J]. Power System Protection and Control, 2026, 54(19).
高伟, 姜一涛, 李奕锋. 基于改进卷积变分自编码器的小样本配电变压器故障诊断方法[J]. 电力系统保护与控制, 2026,54(19). DOI： 10.19783/j.cnki.pspc.260296.

GAO Wei, JIANG Yitao, LI Yifeng. Small-sample distribution transformer fault diagnosis based on improved convolutional variational autoencoder[J]. Power System Protection and Control, 2026, 54(19). DOI： 10.19783/j.cnki.pspc.260296.

摘要

针对配电变压器故障诊断中样本稀缺及未知类别识别困难的问题，提出一种融合改进卷积变分自编码器(improved convolutional variational autoencoder

IM-Conv VAE)与包含概率的支持向量机(probability of inclusion support vector machine

PISVM)的小样本与未知类联合识别方法。首先，通过IM‑Conv VAE对小样本数据进行数据增殖，扩充样本数量。然后，在核边缘损失的联合损失函数约束下进行特征提取。最后，通过引入类别概率边界，利用PISVM对提取的特征进行开放集识别，实现对已知类别的分类与未知类别的标记。实验结果表明，所提模型在样本数量有限的条件下，能精准识别已知故障，并有效区分未知故障。进一步减少训练样本时，模型仍保持良好的稳定性与识别能力。该模型的故障识别准确率高，具备较高的工程应用潜力与推广价值。

Abstract

To address the issues of sample scarcity and difficulty in identifying unknown categories in distribution transformer fault diagnosis

this paper proposes a joint small-sample and unknown-class recognition method that integrates an improved convolutional variational autoencoder (IM‑Conv VAE) with a probability of inclusion support vector machine (PISVM). First

IM‑Conv VAE is used to augment small-sample data and expand the sample size. Then

feature extraction is performed under the joint loss function constraints of core margin loss. Finally

PISVM is constructed by introducing class probability boundaries to perform open-set recognition on the extracted features

thereby achieving classification of known categories and marking of unknown categories. Experimental results show that under limited sample conditions

the proposed model can identify known faults with high accuracy and effectively distinguish unknown faults. Even when the number of training samples is further reduced

the model maintains good stability and strong recognition capability. The model exhibits excellent fault recognition performance and has high potential for engineering application and promotion.

关键词

Keywords

references

TIGHTIZ L, NASAB M A, YANG H, et al. An intelligent system based on optimized ANFIS and association rules for power transformer fault diagnosis[J]. ISA transactions, 2020, 103: 63-74.

CAO C, XU B, LI X. Monitoring method on loosened state and deformational fault of transformer winding based on vibration and reactance information[J]. IEEE Access, 2020, 8: 215479-215492.

TIAN B, FAN X, XU Z, et al. Finite element simulation on transformer vibration characteristics under typical mechanical faults[C] // 2022 IEEE 9th International Conference on Power Electronics Systems and Applications (PESA), September 20-22, 2022, Hong Kong, China: 1-4.

HONG K, PAN J, JIN M. Transformer condition monitoring based on load-varied vibration response and GRU neural networks[J]. IEEE Access, 2020, 8: 178685-178694.

LU G, TSANG W C, YIM N H, et al. Interpretable Fault diagnosis for overhead lines with covered conductors: a physics-informed deep learning approach[J]. Protection and Control of Modern Power Systems, 2025, 10(2): 25-39.

CHENG Q, ZHAO Z, TANG C, et al. Diagnostic of transformer winding deformation fault types using continuous wavelet transform of pulse response[J]. Measurement, 2019, 140: 197-206.

NI R, QIU R, JIN Z, et al. Improved empirical wavelet transform (EWT) and its application in non-stationary vibration signal of transformer[J]. Scientific Reports, 2022, 12(1).

SONG Q, SONG Q, XIAO L, et al. State diagnosis of elevator control transformer over vibration signal based on MEA-BP neural network[J]. Shock and Vibration, 2021(1).

YANG Y, LIU H, HAN L, et al. A feature extraction method using VMD and improved envelope spectrum entropy for rolling bearing fault diagnosis[J]. IEEE Sensors Journal, 2023, 23(4): 3848-3858.

WANG J, GUO J. Application of improved symplectic geometry mode decomposition algorithm in evaluation system of transformer winding short-circuit withstand ability[J]. IEEJ Transactions on Electrical and Electronic Engineering, 2021, 16(10): 1310-1318.

JIANG P, ZHANG Z, ZHANG J, et al. Research on vibration characteristics and multi-parameter state recognition of ±500 kV converter transformer under fluctuating conditions[J]. International Journal of Electrical Power Energy Systems, 2022, 136.

LI C, CHEN J, YANG C, et al. Convolutional neural network-based transformer fault diagnosis using vibration signals[J]. Sensors, 2023, 23(10): 4781.

LI Z, SUN Y, YANG L, et al. Unsupervised machine anomaly detection using autoencoder and temporal convolutional network[J]. IEEE Transactions on Instrumentation and Measurement, 2022, 71: 1-13.

LIAN Y, WANG J, LI Z, et al. Residual attention guided vision transformer with acoustic-vibration signal feature fusion for cross-domain fault diagnosis[J]. Advanced Engineering Informatics, 2025, 64.

SQUIRES S, PRÜGEL BENNETT A, NIRANJAN M. A variational autoencoder for probabilistic non-negative matrix factorisation[J]. Pattern Analysis and Applications, 2025, 28(4): 1-10.

JIANG N, MA N. The establishment and philosophical implications of Hausdorff measure[J]. Afrika Matematika, 2026, 37(2).

SCHEIRER W J, JAIN L P, BOULT T E. Probability models for open set recognition[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2014, 36(11): 2317-2324.

PICKANDS III J. Statistical inference using extreme order statistics[J]. The Annals of Statistics, 1975: 119-131.

LIU H, YANG K, QIAN G, et al. Gramian angular field- Gramian adversial network-ResNet34: high-accuracy fault diagnosis for transformer windings with limited samples[J]. Energies, 2025, 18(16).

LIU S, CHEN J, SHI Z, et al. Representations aligned counterfactual domain learning for open-set fault diagnosis under speed transient conditions[J]. Knowledge-Based Systems, 2025, 310.

XU Y, MA X, WANG X, et al. Unified feature learning network for few-shot fault diagnosis[J]. Neurocomputing, 2024, 598.

HAN G, XIE Y, WANG Z, et al. An open-set classification method with small samples for rotating machinery[J]. IEEE Transactions on Instrumentation and Measurement, 2024, 73: 1-13.

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于小波变换的配电变压器差动保护相位补偿方法

10 kV大容量配电变压器差动保护配置应用研究

基于实时监测数据的海上风电分钟级超短期多步功率预测模型

同塔不换位线路的不对称度对差动保护的影响分析

考虑转供率和容载比的主动配电网供电能力提升研究