基于群组分布鲁棒优化改进XGBoost的变压器故障诊断方法

邹阳; 傅楷宇; 郑智城; 刘凯

doi:10.19783/j.cnki.pspc.260062

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 >

基于群组分布鲁棒优化改进XGBoost的变压器故障诊断方法

应用研究 | 更新时间：2026-07-19

- 基于群组分布鲁棒优化改进XGBoost的变压器故障诊断方法
- Transformer fault diagnosis method based on XGBoost enhanced by group distributionally robust optimization
- 电力系统保护与控制 2026年54卷第14期
- 作者机构：
  
  福州大学电气工程与自动化学院,福建,福州,350108
- 作者简介：
- 基金信息：
- DOI：10.19783/j.cnki.pspc.260062
  中图分类号：
- 纸质出版：2026
- 稿件说明：
移动端阅览
邹阳, 傅楷宇, 郑智城, 等. 基于群组分布鲁棒优化改进XGBoost的变压器故障诊断方法[J]. 电力系统保护与控制, 2026,54(14).

ZOU Yang, FU Kaiyu, ZHENG Zhicheng, et al. Transformer fault diagnosis method based on XGBoost enhanced by group distributionally robust optimization[J]. Power System Protection and Control, 2026, 54(14).
邹阳, 傅楷宇, 郑智城, 等. 基于群组分布鲁棒优化改进XGBoost的变压器故障诊断方法[J]. 电力系统保护与控制, 2026,54(14). DOI： 10.19783/j.cnki.pspc.260062.

ZOU Yang, FU Kaiyu, ZHENG Zhicheng, et al. Transformer fault diagnosis method based on XGBoost enhanced by group distributionally robust optimization[J]. Power System Protection and Control, 2026, 54(14). DOI： 10.19783/j.cnki.pspc.260062.

摘要

在变压器故障诊断中，处于各运行状态分类边缘的样本通常数量稀缺、分布稀疏且存在重叠，导致诊断准确率显著低于基准水平。针对该问题，提出一种基于群组分布鲁棒优化(group distributionally robust optimization

GDRO)改进XGBoost的变压器故障诊断方法。首先，通过高斯混合模型(Gaussian mixture model

GMM)聚类，按油中气体组分差异将各运行状态样本拆分为子类，用以区分集中区与边缘区样本。其次，构建XGBoost故障诊断模型，并采用GDRO算法重新分配各子类权重，以解决模型对边缘样本关注不足的问题。最后，通过算例分析对所提方法的性能进行评估。实验结果表明，相较于常规诊断方法，所提方法对边缘区样本的诊断准确率可达91.67%，总体诊断准确率达到91.08%，显著提升了变压器故障诊断模型在边缘样本上的鲁棒性与全面性。

Abstract

In transformer fault diagnosis

samples located at the classification boundaries of different operating conditions are often scarce

sparsely distributed

and partially overlapping

resulting in significant lower diagnostic accuracy than that achieved on central samples. To address this issue

this paper proposes a transformer fault diagnosis method based on XGBoost enhanced by group distributionally robust optimization (GDRO). First

Gaussian mixture model (GMM) clustering is used to cluster the samples and divide them into subclasses based on the differences in gas components in the oil

distinguishing between the central and boundary region samples. Then

an XGBoost fault diagnosis model is constructed

and GDRO algorithm is adopted to reassign weights to different subclasses

overcoming the model's tendency to underemphasize boundary samples. Finally

case studies are conducted to evaluate the performance of the proposed method. Experimental results demonstrate that

compared to conventional diagnostic methods

the proposed approach achieves a diagnostic accuracy of 91.67% for samples in boundary regions and an overall diagnostic accuracy of 91.08%

significantly enhancing the robustness and comprehensiveness of the transformer fault diagnosis model with respect to boundary samples.

关键词

Keywords

references

HEMMATIAN M, GHANBARI H, KHOSRAVI A. Addressing imbalanced data classification with cluster- based reduced noise SMOTE[J]. PLoS ONE, 2025, 15(6).

TAN J J, KWAN B H, NG D W K, et al. Adaptive focal loss with personality stratification for stably mitigating hard class imbalance in multi‑dimensional personality recognition[J]. Scientific Reports, 2025, 15.

CHUNG Y, LEE H J. Joint triplet loss with semi-hard constraint for data augmentation and disease prediction using gene expression data[J]. Scientific Reports, 2023, 13.

SOHONI N S, DUNNMON J A, ANGUS G, et al. No subclass left behind: fine-grained robustness in coarse- grained classification problems[C] // Proceedings of the 34th Conference on Neural Information Processing Systems (NIPS), December 6-12, 2020, Vancouver BC, Canada: 19339-19352.

BEN-TAL A, DEN HERTOG D, DE WAEGENAERE A M B, et al. Robust solutions of optimization problems affected by uncertain probabilities[J]. Management Science, 2013, 59(2): 341-357.

SAGAWA S, KOH P W, HASHIMOTO T B, et al. Distributionally robust neural networks for group shifts: on the importance of regularization for worst‑case generalization[C] // International Conference on Learning Representations (ICLR), April 30, 2020, Addis Ababa, Ethiopia: 2661-2679.

THUDI A, MADDISON C J. MixMax: distributional robustness in function space via optimal data mixtures[C] // International Conference on Learning Representations (ICLR), April 24-28, 2025, Singapore, Singapore: 28975-28995.

DUVAL M, DEPABLA A. Interpretation of gas-in-oil analysis using new IEC publication 60599 and IEC TC 10 databases[J]. IEEE Electrical Insulation Magazine, 2001, 17(2): 31-41.

孙才新, 陈伟根, 李俭, 等. 电气设备油中气体在线监测与故障诊断技术[M]. 北京: 科学出版社, 2003.

蔡金锭, 邹阳. 电力变压器智能故障诊断与绝缘测试技术[M]. 北京: 电子工业出版社, 2017.

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

基于环面积图谱熵特征的变压器局部放电故障诊断

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

基于天牛须搜索优化支持向量机的变压器故障诊断研究

基于 CVaR 约束的双层主从博弈优化模型及其在多政策场景下的应用研究