A cloud-edge collaborative mobile AI inspection closed-loop management method and system based on knowledge graph enhancement

By constructing a physical topology degradation propagation map of equipment and employing cloud-edge collaboration, the problem of insufficient root cause tracing capability in mobile AI inspection systems was solved, enabling accurate root cause localization and prediction of equipment anomalies and improving the efficiency and accuracy of the inspection system.

CN122839283APending Publication Date: 2026-09-29BEIJING GRAPHSAFE TECH CO LTD
0 Cites 0 Cited by

Patent Information

Application Number
CN202611087498.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-07-22
Publication Date
2026-09-29

AI Technical Summary

Technical Problem

Existing mobile AI inspection systems lack physical causal semantics and root cause tracing capabilities in equipment anomaly detection, causing maintenance personnel to perform only superficial maintenance operations on symptom-related components, leading to frequent recurrence of problems.

Method used

Construct a physical topology degradation propagation map of the equipment, configure propagation weights and visual features of health status, utilize edge computing and cloud collaboration, and use graph variational autoencoders and causal masking matrices to detect equipment anomalies and locate root causes, generate maintenance task packages and perform targeted re-sampling verification, and update the map propagation weights.

Benefits of technology

It improved the accuracy of root cause localization of potential hazards, reduced false correlation reasoning errors, and enabled effective root cause localization and prediction of equipment anomalies, thereby improving inspection efficiency and accuracy.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The application provides a cloud-edge collaborative mobile AI inspection closed-loop management method and system based on knowledge graph enhancement, and relates to the technical field of intelligent operation and maintenance of industrial equipment, wherein the method comprises: taking equipment components as nodes and physical connections as directed edges to construct a degradation propagation graph, and configuring propagation weights, visual features and coordinates; an edge gateway generates a real-time state subgraph by processing inspection multi-modal data based on the graph, and discriminates equipment abnormalities and reports deviation topology by means of a graph variational autoencoder; a cloud end constructs a causal shielding matrix in combination with the graph propagation weights, constrains reasoning along the physical degradation direction by fusing a large model attention matrix, realizes root cause positioning and propagation path prediction; and according to this, maintenance and targeted resampling instructions are issued, and after resampling verification, the graph triplets and propagation weights are updated to form a closed loop. The application combines physical priori and data driving, avoids the problem of false association of pure data models, and significantly improves the accuracy of equipment hidden danger root cause positioning.
Need to check novelty before this filing date? Find Prior Art