Dynamic identification method, device, electronic equipment and storage medium for voltage control operation conditions in new power distribution networks

By combining a self-attention neural network architecture with cross-attention and multi-head self-attention mechanisms, the dynamic adaptability and collaborative control problems of distribution network voltage control are solved, enabling rapid and accurate identification of distribution network voltage operating conditions and efficient voltage control.

CN121834474BActive Publication Date: 2026-05-26JINCHENG POWER SUPPLY COMPANY OF STATE GRID SHANXI ELECTRIC POWER
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
JINCHENG POWER SUPPLY COMPANY OF STATE GRID SHANXI ELECTRIC POWER
Filing Date
2026-03-11
Publication Date
2026-05-26

AI Technical Summary

Technical Problem

Existing voltage control methods for distribution networks cannot meet the dynamic adaptability and coordinated control requirements of new distribution networks, leading to voltage control lag, misjudgment, and grid stability issues.

Method used

A self-attention-based neural network architecture is adopted, which combines cross-attention and multi-head self-attention mechanisms to construct electrical features and perform feature fusion and extraction. Electrical information of voltage operating condition categories is generated through unsupervised classification.

Benefits of technology

It enables rapid and accurate dynamic identification of voltage operating conditions in the power distribution network, improves the dynamic response capability of the power grid and the scientificity and accuracy of voltage control, and adapts to the development trend of modern power distribution systems.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121834474B_ABST
    Figure CN121834474B_ABST
Patent Text Reader

Abstract

This application provides a novel dynamic discrimination method, device, electronic equipment, and storage medium for voltage control operation conditions in distribution networks, relating to the field of power system research. The discrimination method includes: constructing a first electrical feature quantity and a second electrical feature quantity; inputting the two types of electrical feature quantities into a self-attention-based neural network architecture; fusing the two types of electrical feature quantities using a cross-attention mechanism; extracting the fused features using a multi-head self-attention mechanism; performing pooling operations on the extracted features; inputting the pooled features into a fully connected neural network to parse the interrelationships between the features; inputting the feature relationships into a classifier for unsupervised classification to generate feature sequences corresponding to each operation condition category; and decoding the feature sequences to generate electrical information quantities corresponding to each operation condition category. This scheme fully considers the control potential of multi-level substations and new energy sources, generating multiple categories of operation conditions, and providing efficient and accurate guidance for distribution network voltage control decisions.
Need to check novelty before this filing date? Find Prior Art