Method and device for identifying arc discharge mode in short-gap oil of transformer

Through the combined detection of multi-physical quantities and the deep-separable convolutional neural network model, the problem of low accuracy of single physical signal recognition in arc discharge fault diagnosis is solved, and higher recognition accuracy and robustness are achieved to ensure the safety of power equipment.

CN119044689AActive Publication Date: 2024-11-29XI AN JIAOTONG UNIV
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
CN202411106245.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-13
Publication Date
2024-11-29
Estimated Expiration
2044-08-13

AI Technical Summary

Technical Problem

The prior art relies on a single physical signal in arc discharge fault diagnosis, is susceptible to noise interference, has low recognition accuracy, and is difficult to fully cope with all types of faults, especially in the short gap arc discharge stage, posing a threat to the safety of power equipment.

Method used

Collect multi-physical signals of arc discharge (current, ultrasound, ultra-high frequency, oil pressure, etc.), use generalized S transformation to generate time-frequency maps, and jointly detect them through multi-channel depth separable convolutional neural network model to realize discharge pattern recognition .

Benefits of technology

The accuracy and robustness of discharge pattern recognition are improved, and the problems of insufficient information and low fault tolerance of single physical quantity recognition are avoided. The recognition accuracy reaches 98.730%, effectively preventing the development of short gap arc discharge to a more serious stage.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119044689A_ABST
    Figure CN119044689A_ABST
Patent Text Reader

Abstract

The invention discloses a method and a device for identifying an arc discharge mode in short-gap oil of a transformer, and the method comprises the steps: carrying out an oil-paper insulation arc discharge experiment based on different discharge defect models of an oil-immersed power transformer; a sensor is used for collecting multi-physical characteristic signals of arc discharge in short-gap oil in an oil-paper insulation arc discharge experiment; processing the multi-physical characteristic signal by using generalized S-transform to generate a multi-physical signal time-frequency map; and inputting the generated multi-physical-signal time-frequency atlas into a multi-channel depth separable convolutional neural network model to realize arc discharge mode identification of multi-physical-quantity joint detection.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Insulation defect detection experiment platform and method under oil-immersed power equipment working condition simulation

    CN112462209A

  • Transformer partial discharge fault mode identification method and system

    CN116340812A

  • Method and device for diagnosing fault arc, and computer-readable storage medium

    WO2022067562A1