一种基于几何特征映射与增强型EEGNet的有载分接开关故障诊断方法
By using a method based on geometric feature mapping and enhanced EEGNet, the problems of insufficient signal topology information mining and opaque deep learning models in on-load tap changer fault diagnosis are solved, enabling accurate capture and reliable diagnosis of weak faults and improving the reliability of power equipment operation and maintenance.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SHANDONG UNIV
- Filing Date
- 2026-05-09
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies for fault diagnosis of on-load tap changers suffer from insufficient mining of spatial topological information of signal feature evolution, poor feature robustness under strong background noise, and opaque decision-making of deep learning models, making it difficult to achieve accurate capture and reliable diagnosis of weak faults.
A fault diagnosis method based on geometric feature mapping and enhanced EEGNet is adopted, including Hilbert-Savgol filtering, phase space mapping, dual-channel topology map construction and improved EEGNet network. Signal features are extracted through multi-scale convolution and adaptive attention mechanism, and SHAP attribution analysis method is introduced for interpretation.
It effectively improves the accuracy and transparency of identifying on-load tap changer faults, provides intuitive diagnostic evidence, and enhances the credibility of power equipment operation and maintenance.
Smart Images
Figure CN122153755B_ABST