Energy storage system demand response prediction method and system based on deep learning

By constructing a multi-dimensional feature matrix and risk situation description using deep learning technology, and combining it with real-time adjustment resource data, the problem of insufficient accuracy in traditional energy storage system demand response prediction methods has been solved. This enables precise location and efficient adjustment of weak links in the power grid, thereby improving the frequency stability and operational safety of the power grid.

CN121840561APending Publication Date: 2026-04-10STATE GRID HUBEI ELECTRIC POWER RES INST
View PDF 0 Cites 1 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-12-02
Publication Date
2026-04-10

AI Technical Summary

Technical Problem

Traditional demand response forecasting methods for energy storage systems cannot accurately reflect the complexity of the power grid operating environment, resulting in low accuracy of demand response forecasts. This makes it difficult to effectively cope with the strong fluctuations and uncertainties in the output power of wind turbines, affecting the stability of power grid frequency and operational safety.

Method used

By acquiring historical operation data and grid status data associated with the energy storage system, a multi-dimensional feature matrix is ​​constructed using deep learning technology. Combined with environmental prediction parameters and regulation resource data, collaborative analysis and risk situation description are performed to accurately locate weak links in the grid, provide forward-looking risk situation information, and dynamically correct the situation by combining real-time regulation resources to generate accurate demand response instructions.

Benefits of technology

It significantly improves the accuracy and feasibility of demand response forecasting for energy storage systems, enabling early identification of periods with high forecast uncertainty, precise location of weak links in the power grid, ensuring that the regulation capacity of the energy storage system matches the actual demand of the power grid, and enhancing the power grid's ability to cope with power fluctuations.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121840561A_ABST
    Figure CN121840561A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of power grid systems, and discloses an energy storage system demand response prediction method and system based on deep learning. According to the invention, through analyzing the matching relationship between the statistical characteristics of historical prediction errors and the meteorological mode, early warning time periods which are high in prediction uncertainty and need to be focused by an energy storage system can be identified through the obtained power prediction credibility time sequence, and through obtaining the power transmission risk index sequence, the early warning time periods can be identified. Therefore, the uncertainty of power prediction is converted into a specific power grid security risk index to provide prospective risk situation information, and the dynamic coupling relationship among the power supply uncertainty, the power transmission bottleneck and the load demand is quantified by constructing a multi-dimensional feature matrix to improve the power supply reliability. Therefore, a mapping bridge from the operation state of the power grid to the energy storage response demand is established, the problem that the prediction result is disjointed with the actual safety demand of the power grid due to the single prediction dimension in the traditional method is solved, and the accuracy and the performability of energy storage demand response prediction can be remarkably improved.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Cited By

  • An ultra-short-term offshore wind power prediction method considering multi-wind farm information

    CN122203224A