An MPC and AGNN-GC-based electric heating optimization control method

By combining MPC and AGNN-GC methods, the accuracy of electric heating control strategies and the intensive management of user energy consumption were improved, solving the problems of low resource utilization and extensive user management in electric heating systems, and promoting the consumption of new energy sources and load stability.

CN117212881BActive Publication Date: 2026-07-21NANJING UNIV OF POSTS & TELECOMM
2 Cites 0 Cited by

Patent Information

Application Number
CN202311242091.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-09-25
Publication Date
2026-07-21
Estimated Expiration
2043-09-25

AI Technical Summary

Technical Problem

Electric heating suffers from problems such as inadequate supporting dispatching methods, low utilization rate of electric heating resources, and extensive management of heating users. Furthermore, the potential for independent electric heating users to participate in demand response has not been fully explored.

Method used

The method employs Model Predictive Control (MPC) algorithm and Adaptive Graph Neural Network Graph Convolutional Control (AGNN-GC) method, combining feature groups of MPC algorithm, natural feature groups, human behavioral feature groups and building feature groups. The electric heating control strategy is globally and locally embedded through graph convolutional network module and graph attention network module to generate real-time operation strategy.

Benefits of technology

It has improved the accuracy of electric heating control strategies, promoted the consumption of new energy sources, realized the intensive management of energy consumption for electric heating users, reduced heating costs, and smoothed load fluctuations.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117212881B_ABST
    Figure CN117212881B_ABST
Patent Text Reader

Abstract

The application discloses an electric heating optimization control method based on MPC and AGNN-GC, and belongs to the technical field of electric heating control. The method comprises the following steps: acquiring electric heating control strategies of factor groups with different characteristics; constructing a thermal dynamics model, and establishing a discrete state space model through the constructed thermal dynamics model; adopting an MPC algorithm to optimize and solve the electric heating control strategies of the model prediction control algorithm characteristic factor groups; constructing an AGNN-GC network model, and fusing the electric heating control strategies of the factor groups with different characteristics by using the AGNN-GC network model; adopting a focal loss to balance positive and negative samples, training the AGNN-GC network model, and generating a real-time operation strategy of electric heating. The application adopts the MPC algorithm to solve the offline control strategies of the electric heating of the model prediction control algorithm characteristic factor groups, and then adopts the AGNN-GC model to globally and locally embed the operation power in all factor groups, so that the accuracy of the generated real-time operation strategy of electric heating is improved, and the energy interaction of the electric heating system is optimized.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Heat accumulating type electric heating power utilization control decision method and system based on Pso-CNN

    CN114742410A

  • Heat supply system control method based on self-adaptive model predictive control technology

    CN116734307A