JOURNAL ARTICLE

Research and application of transformer fault diagnosis method based on SHAP-IHHO-Xgboost model.

  • Published In: Journal of Computational Methods in Sciences & Engineering (Sage Publications Inc.), 2025, v. 25, n. 6. P. 5715 1 of 3

  • Database: Academic Search Ultimate 2 of 3

  • Authored By: Liu, Cheng 3 of 3

Abstract

The article focuses on developing a hybrid fault diagnosis model for power transformers to improve diagnostic accuracy, feature utilization, and model interpretability. It proposes combining SHapley Additive exPlanations (SHAP) for feature selection, an improved Harris Hawks optimization (IHHO) algorithm for hyperparameter tuning, and the eXtreme gradient boosting (Xgboost) model for classification. Experimental results using dissolved gas analysis (DGA) data from transformers demonstrate that the SHAP-IHHO-Xgboost model achieves superior performance, with an accuracy of 0.9509, outperforming traditional methods and other optimization algorithms. The study also discusses the model's limitations, including lower accuracy for certain fault categories, computational complexity, and reliance solely on DGA data, suggesting future work on multimodal data integration and real-time diagnosis enhancements.

Additional Information

  • Source:Journal of Computational Methods in Sciences & Engineering (Sage Publications Inc.). 2025/11, Vol. 25, Issue 6, p5715
  • Document Type:Article
  • Subject Area:Engineering
  • Publication Date:2025
  • ISSN:1472-7978
  • DOI:10.1177/14727978251352157
  • Accession Number:188762415
  • Copyright Statement:Copyright of Journal of Computational Methods in Sciences & Engineering (Sage Publications Inc.) is the property of Sage Publications Inc. and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Looking to go deeper into this topic? Look for more articles on EBSCOhost.