GIS insulation defect diagnosis method and system

By deeply fusioning acoustic, optical, and electronic signals and deploying a lightweight multimodal large model, combined with a confidence-driven cloud-edge collaborative architecture, the problems of insufficient multimodal collaboration and poor real-time performance at the edge are solved in GIS insulation defect diagnosis, achieving high-precision and real-time diagnostic results.

CN122449339APending Publication Date: 2026-07-24蔡昕霖
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
蔡昕霖
Filing Date
2026-04-30
Publication Date
2026-07-24

Smart Images

  • Figure CN122449339A_ABST
    Figure CN122449339A_ABST
Patent Text Reader

Abstract

The application discloses a GIS insulation defect diagnosis method and system, collects ultrasonic waves, light pulses and ultra-high frequency signals, and carries out denoising and space-time alignment, forms unified representation after modality coding and cross attention alignment of the three modal signals, inputs an edge lightweight multi-modal model to obtain preliminary diagnosis results and confidence, and when the confidence is lower than a threshold value, triggers cloud edge cooperation, completes deep reasoning by a cloud teacher model and outputs an explanatory report. The application is suitable for GIS equipment online insulation defect diagnosis.
Need to check novelty before this filing date? Find Prior Art