A transformer inter-turn short circuit fault diagnosis method and system based on broadband current

By employing wideband current multi-dimensional analysis and intelligent diagnostic algorithms, the problem of real-time and accurate diagnosis of inter-turn short-circuit faults in transformers has been solved, improving the ability to identify and locate early faults and making it suitable for online monitoring of dry-type transformers.

CN122131193APending Publication Date: 2026-06-02CHINA YANGTZE POWER
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
CHINA YANGTZE POWER
Filing Date
2026-03-17
Publication Date
2026-06-02

AI Technical Summary

Technical Problem

Existing methods for diagnosing transformer inter-turn short-circuit faults are not sensitive enough, are susceptible to environmental noise interference, and cannot be monitored in real time, making it difficult to meet the power system's need for accurate diagnosis of early faults.

Method used

Based on a multi-dimensional analytical method using broadband current, combined with empirical mode decomposition (EMD) and Hilbert spectral analysis, a random forest model optimized by particle swarm optimization is used for fault severity classification and diagnosis, and a weighted Mahalanobis distance-fault location correlation algorithm is used to achieve accurate location.

Benefits of technology

It enables real-time and sensitive diagnosis of inter-turn short-circuit faults in transformers, accurately identifies weak faults, has strong anti-interference capabilities, and high positioning accuracy. It is suitable for dry-type transformers of different capacities and types.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122131193A_ABST
    Figure CN122131193A_ABST
Patent Text Reader

Abstract

This invention provides a method and system for diagnosing transformer inter-turn short-circuit faults based on broadband current, belonging to the field of power equipment fault diagnosis technology. The system includes a signal acquisition module, a feature extraction module, a fault severity diagnosis module, and a fault location module. The signal acquisition module captures and preprocesses the broadband current signal at the transformer core grounding wire; the feature extraction module extracts multi-dimensional features such as peak current, equivalent duration, 0-2MHz energy proportion, and dominant frequency; the fault severity diagnosis module uses a random forest model optimized by particle swarm optimization to achieve three levels of fault classification: mild, moderate, and severe; the fault location module constructs a simulated feature dataset and combines it with a weighted Mahalanobis distance-fault location association algorithm to achieve precise fault location. This invention does not require interrupting transformer operation, allows for real-time monitoring, has high diagnostic sensitivity and strong anti-interference capability, and a location accuracy rate of over 90%. It is applicable to dry-type transformers of different capacities and can effectively ensure the safe and stable operation of the power system.
Need to check novelty before this filing date? Find Prior Art