An interpretable machine learning method for fault diagnosis of oil-immersed transformers based on edge inference

Intelligent diagnostic models using dissolved gas analysis are crucial for oil-immersed transformer fault diagnosis. However, the inherent “black box” nature of these models limits interpretability, and traditional methods that upload local data to central servers raise data security concerns. To ad...

Full description

Saved in:
Bibliographic Details
Main Authors: Dong Hu, Yong Yang, Hao Dai, Chao Tang, Jufang Xie
Format: Article
Language:English
Published: Elsevier 2025-07-01
Series:International Journal of Electrical Power & Energy Systems
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S014206152500198X
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850272190411309056
author Dong Hu
Yong Yang
Hao Dai
Chao Tang
Jufang Xie
author_facet Dong Hu
Yong Yang
Hao Dai
Chao Tang
Jufang Xie
author_sort Dong Hu
collection DOAJ
description Intelligent diagnostic models using dissolved gas analysis are crucial for oil-immersed transformer fault diagnosis. However, the inherent “black box” nature of these models limits interpretability, and traditional methods that upload local data to central servers raise data security concerns. To address these issues, this study proposes an interpretable fault diagnosis model for edge deployment. First, a filtered feature extraction algorithm based on real domain rough set theory is proposed to optimize feature extraction before model input. Experimental results demonstrate that this algorithm enhances model performance and reduces inference time at the edge-end. Second, the hyperparameters of Extreme Gradient Boosting are automatically tuned using the Newton–Raphson optimizer. Compared with other diagnostic methods, the proposed model yields superior classification effect accuracy. Following edge-end inference, the SHapley Additive exPlanations method is employed to analyze feature impact on diagnostic results, visualizing the significance of different characteristic gases for fault types using SHAP values. Finally, the model’s robustness, reliability, and interpretability are validated through real cases, providing practical insights for transformer operation and maintenance.
format Article
id doaj-art-98fdf9cd4dce4968a736701871b54511
institution OA Journals
issn 0142-0615
language English
publishDate 2025-07-01
publisher Elsevier
record_format Article
series International Journal of Electrical Power & Energy Systems
spelling doaj-art-98fdf9cd4dce4968a736701871b545112025-08-20T01:51:54ZengElsevierInternational Journal of Electrical Power & Energy Systems0142-06152025-07-0116811064710.1016/j.ijepes.2025.110647An interpretable machine learning method for fault diagnosis of oil-immersed transformers based on edge inferenceDong Hu0Yong Yang1Hao Dai2Chao Tang3Jufang Xie4College of Engineering and Technology, Southwest University, Chongqing, China; International R&D Center for New Technologies of Smart Grid and Equipment, Southwest University, Chongqing, China; Corresponding author at: College of Engineering and Technology, Southwest University, Chongqing, China.College of Engineering and Technology, Southwest University, Chongqing, China; International R&D Center for New Technologies of Smart Grid and Equipment, Southwest University, Chongqing, ChinaCollege of Engineering and Technology, Southwest University, Chongqing, China; International R&D Center for New Technologies of Smart Grid and Equipment, Southwest University, Chongqing, ChinaCollege of Engineering and Technology, Southwest University, Chongqing, China; International R&D Center for New Technologies of Smart Grid and Equipment, Southwest University, Chongqing, ChinaCollege of Engineering and Technology, Southwest University, Chongqing, China; International R&D Center for New Technologies of Smart Grid and Equipment, Southwest University, Chongqing, ChinaIntelligent diagnostic models using dissolved gas analysis are crucial for oil-immersed transformer fault diagnosis. However, the inherent “black box” nature of these models limits interpretability, and traditional methods that upload local data to central servers raise data security concerns. To address these issues, this study proposes an interpretable fault diagnosis model for edge deployment. First, a filtered feature extraction algorithm based on real domain rough set theory is proposed to optimize feature extraction before model input. Experimental results demonstrate that this algorithm enhances model performance and reduces inference time at the edge-end. Second, the hyperparameters of Extreme Gradient Boosting are automatically tuned using the Newton–Raphson optimizer. Compared with other diagnostic methods, the proposed model yields superior classification effect accuracy. Following edge-end inference, the SHapley Additive exPlanations method is employed to analyze feature impact on diagnostic results, visualizing the significance of different characteristic gases for fault types using SHAP values. Finally, the model’s robustness, reliability, and interpretability are validated through real cases, providing practical insights for transformer operation and maintenance.http://www.sciencedirect.com/science/article/pii/S014206152500198XDissolved gas analysisExplainable machine learningEdge-endFeature selectionGradient boostingIntelligent diagnostic models
spellingShingle Dong Hu
Yong Yang
Hao Dai
Chao Tang
Jufang Xie
An interpretable machine learning method for fault diagnosis of oil-immersed transformers based on edge inference
International Journal of Electrical Power & Energy Systems
Dissolved gas analysis
Explainable machine learning
Edge-end
Feature selection
Gradient boosting
Intelligent diagnostic models
title An interpretable machine learning method for fault diagnosis of oil-immersed transformers based on edge inference
title_full An interpretable machine learning method for fault diagnosis of oil-immersed transformers based on edge inference
title_fullStr An interpretable machine learning method for fault diagnosis of oil-immersed transformers based on edge inference
title_full_unstemmed An interpretable machine learning method for fault diagnosis of oil-immersed transformers based on edge inference
title_short An interpretable machine learning method for fault diagnosis of oil-immersed transformers based on edge inference
title_sort interpretable machine learning method for fault diagnosis of oil immersed transformers based on edge inference
topic Dissolved gas analysis
Explainable machine learning
Edge-end
Feature selection
Gradient boosting
Intelligent diagnostic models
url http://www.sciencedirect.com/science/article/pii/S014206152500198X
work_keys_str_mv AT donghu aninterpretablemachinelearningmethodforfaultdiagnosisofoilimmersedtransformersbasedonedgeinference
AT yongyang aninterpretablemachinelearningmethodforfaultdiagnosisofoilimmersedtransformersbasedonedgeinference
AT haodai aninterpretablemachinelearningmethodforfaultdiagnosisofoilimmersedtransformersbasedonedgeinference
AT chaotang aninterpretablemachinelearningmethodforfaultdiagnosisofoilimmersedtransformersbasedonedgeinference
AT jufangxie aninterpretablemachinelearningmethodforfaultdiagnosisofoilimmersedtransformersbasedonedgeinference
AT donghu interpretablemachinelearningmethodforfaultdiagnosisofoilimmersedtransformersbasedonedgeinference
AT yongyang interpretablemachinelearningmethodforfaultdiagnosisofoilimmersedtransformersbasedonedgeinference
AT haodai interpretablemachinelearningmethodforfaultdiagnosisofoilimmersedtransformersbasedonedgeinference
AT chaotang interpretablemachinelearningmethodforfaultdiagnosisofoilimmersedtransformersbasedonedgeinference
AT jufangxie interpretablemachinelearningmethodforfaultdiagnosisofoilimmersedtransformersbasedonedgeinference