A method and system for detecting defects in oil-paper insulation of an oil-immersed transformer

CN122289249APending Publication Date: 2026-06-26CHINA UNIV OF MINING & TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202610602776.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-05-06
Publication Date
2026-06-26

AI Technical Summary

Technical Problem

Existing technologies are insufficient to accurately detect the location and severity of defects in the internal oil-paper insulation of oil-immersed transformers. Furthermore, traditional detection methods are prone to damaging the insulation environment, have long inspection cycles, and are costly.

Method used

A miniature vision robot is used in conjunction with an anti-signal-shielding ultrasonic sensor and a vision sensor for detection. By constructing a defect intelligent recognition model, and utilizing a decoder with bright channel prior theory, differential parallel feature convolution, sparse attention fusion skip connections, and weighted calibration channel composite stitching, image enhancement and defect segmentation are achieved.

Benefits of technology

It improves the accuracy and efficiency of detecting defects in oil-paper insulation, shortens the inspection cycle, reduces engineering costs, and provides a guarantee for the safe operation of the power grid.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122289249A_ABST
    Figure CN122289249A_ABST
Patent Text Reader

Abstract

This invention relates to the field of transformer defect analysis and detection technology, specifically a method and system for detecting defects in the oil-paper insulation of oil-immersed transformers. Addressing the problems of poor image quality and low defect recognition accuracy in images acquired by underwater vehicles inside transformers, this invention first enhances insulation defect images using a bright channel prior algorithm with color weight correction. Then, it constructs an intelligent defect recognition model consisting of an encoder with differential parallel feature convolution, a skip connection with sparse attention fusion, and a decoder with weighted calibrated channel composite splicing. After training with a carbon trace defect dataset, the model detects images acquired in real-time by the underwater vehicle, solving problems of uneven image exposure and color distortion, accurately extracting carbon trace defect features, and achieving precise positioning of irregular edges. This improves defect segmentation accuracy and detection efficiency, enabling efficient monitoring of the transformer's oil-paper insulation status and providing a guarantee for the safe operation of the power grid.
Need to check novelty before this filing date? Find Prior Art