Fusion processing method, system and equipment for multi-source heterogeneous data and medium

By constructing a wireless adaptive sensor network and a deep denoising autoencoder combined with a self-attention mechanism and a graph neural network, the problem of fusion of multi-source heterogeneous data and intelligent decision-making in the power Internet of Things was solved, achieving real-time adaptation and efficient data processing, and improving communication reliability and fault early warning accuracy.

CN121397489APending Publication Date: 2026-01-23GUIZHOU POWER GRID CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202511454955.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-10-13
Publication Date
2026-01-23

AI Technical Summary

Technical Problem

Existing technologies suffer from poor protocol compatibility, insufficient dynamic response capabilities, and limited intelligence levels in processing multi-source heterogeneous data in the context of the Internet of Things (IoT) for the power industry. They are unable to effectively integrate structured and unstructured data, and traditional methods cannot dynamically mine potential correlation value, resulting in response delays and insufficient system efficiency of the communication resource management platform.

Method used

A wireless adaptive sensor network is constructed, using a distributed message queue middleware and a deep denoising autoencoder for data buffering and denoising operations. The self-attention mechanism and graph neural network are combined to dynamically calculate data weights, and an elastic weight solidification strategy is adopted to achieve incremental learning and construct a multi-source data graph structure.

Benefits of technology

It achieves real-time adaptation of multi-protocol data fusion, improves communication reliability and fault early warning accuracy, reduces operation and maintenance costs, maintains millisecond-level response capability in terminal-intensive scenarios, and solves the problem of dynamic fusion and intelligent decision-making of multi-source heterogeneous data.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121397489A_ABST
    Figure CN121397489A_ABST
Patent Text Reader

Abstract

The invention discloses a multi-source heterogeneous data fusion processing method, system and device and a medium, and belongs to the technical field of data processing and fusion. A wireless adaptive sensor network is constructed, the data sampling frequency and range are dynamically adjusted according to the environment, and multi-source data are classified, buffered and transmitted through distributed message queue middleware; a deep denoising auto-encoder is adopted to execute data denoising and exception repair, and heterogeneous data is processed in a unified mode in combination with a multi-level standardization mechanism; data source weights are dynamically calculated based on a self-attention mechanism, multi-source data association is modeled in combination with a graph neural network, and incremental learning is realized by adopting an elastic weight solidification strategy; the problem of multi-protocol fusion misalignment is solved through a dynamic standardization processing unit, structural and non-structural data splitting bottlenecks are broken through by associating a hierarchical fusion decision engine with a network topology, performance degradation during system expansion is eliminated by using an incremental resource scheduler, and millisecond-level response capability is maintained in a terminal dense scene.
Need to check novelty before this filing date? Find Prior Art