Power system twin modeling method, device and equipment based on renewable energy grid connection

Through the particle swarm algorithm, the initial mechanism model and long-term and short-term neural network feedback were corrected, and the digital twin model of the power system was constructed, which solved the computational complexity and overfitting problems, and improved the model accuracy and operating efficiency of the power system.

CN120597720APending Publication Date: 2025-09-05GUANGDONG POWER GRID CORP ZHAOQING POWER SUPPLY BUREAU
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202510792500.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-06-13
Publication Date
2025-09-05

AI Technical Summary

Technical Problem

The existing power system simulation model based on long-term and short-term neural networks is complex in the calculation process and has overfitting problems, resulting in a large deviation between the digital twin model and the real physical object data, affecting the accuracy of power system regulation and fault diagnosis.

Method used

The particle swarm algorithm is used to correct the initial mechanism model parameters, combine long-term and short-term neural networks and real-time data feedback to build a digital twin model, and optimize the model parameters through the power error feedback mechanism until the error threshold is reached to ensure the model accuracy.

Benefits of technology

It reduces the computational complexity and training time of long-term and short-term neural networks, improves the real-time accuracy of digital twin models, avoids overfitting, and enhances the operating efficiency and reliability of the power system.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120597720A_ABST
    Figure CN120597720A_ABST
Patent Text Reader

Abstract

The invention relates to a power system twin modeling method, device and equipment based on renewable energy grid connection, and the method comprises the steps: obtaining the power data of each power generation equipment in a power system, and carrying out the construction according to the power data of each power generation equipment and a particle swarm algorithm, and obtaining a mechanism model; acquiring array historical data and real-time data of the power system, and performing power prediction by adopting a long-short-term neural network to obtain predicted power; acquiring a topological structure of the power system and acquisition data of each acquisition point in a set time period, and constructing a digital twinborn model in the power system simulation software according to the topological structure and the mechanism model of each power generation device; controlling the digital twin model to perform simulation operation according to each piece of acquired data to obtain simulation output power; calculating to obtain a power error mean value according to all the simulation output power and the predicted power; whether the digital twinborn model is updated or not is determined according to the power error mean value to obtain the digital twinborn model of the power system, and the model construction calculation strength is reduced.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Cited By

  • Digital twinning-based power grid transient process deduction method, electronic equipment and medium

    CN120768020A

  • Gas turbine power generation system three-dimensional model construction method based on digital twinning

    CN121543409A