A power distribution station instrument intelligent identification and question and answer diagnosis method and system

By extracting deep image features and text semantic features of substation instruments through a joint image-text coding model, the problem of insufficient robustness of instrument recognition in complex environments is solved. This enables the integration of intelligent recognition and question-and-answer diagnosis of substation instruments, improving the accuracy of equipment status judgment and operation and maintenance efficiency.

CN122133804APending Publication Date: 2026-06-02NANJING LINGSHU INTELLIGENT TECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
NANJING LINGSHU INTELLIGENT TECHNOLOGY CO LTD
Filing Date
2026-03-16
Publication Date
2026-06-02

AI Technical Summary

Technical Problem

Existing substation instrument identification technologies lack robustness in complex environments and cannot achieve integrated image recognition and natural language interaction, resulting in inaccurate equipment status judgments and failing to meet the needs of efficient operation and maintenance of modern power distribution networks.

Method used

A graph-text joint coding model is adopted, which extracts the deep image features of instrument image data and the text semantic features of natural language questions by setting up visual coding branch networks and language coding branch networks in parallel. The graph-text semantic alignment is then performed, and semantic reasoning is carried out in combination with the substation operation and maintenance threshold library and knowledge base to output question-answering diagnostic results.

Benefits of technology

It improves the robustness of instrument identification in complex environments, supports integrated equipment status judgment and diagnosis, and meets the real-time and intelligent operation and maintenance needs of modern power distribution networks.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122133804A_ABST
    Figure CN122133804A_ABST
Patent Text Reader

Abstract

This invention provides a method and system for intelligent identification and question-and-answer diagnosis of instruments in power distribution stations, comprising: acquiring instrument image data collected on-site in the power distribution station and natural language questions input by the user; based on the instrument image data and the natural language questions, using a graph-text joint coding model, outputting the deep image features of the instrument image data and the text semantic features of the natural language questions; performing graph-text semantic alignment on the deep image features and the text semantic features to obtain graph-text fusion features; obtaining the instrument reading extraction result based on the graph-text fusion features; performing semantic reasoning based on the instrument reading extraction result, the natural language questions, the operation and maintenance threshold library, and the operation and maintenance knowledge base, and outputting a question-and-answer diagnosis result matching the natural language questions; the method of this invention can realize human-computer interactive intelligent question-and-answer diagnosis function, and effectively improve the robustness of instrument identification and the intelligence level of operation and maintenance diagnosis under complex operating conditions, thereby simplifying the interaction and decision-making process of intelligent inspection of power distribution stations.
Need to check novelty before this filing date? Find Prior Art