Energy Management Method for Electro-Thermal Coupled New Energy Systems Based on Deep Reinforcement Learning

By developing an energy management method for electro-thermal coupled new energy systems based on deep reinforcement learning, we have solved the limitations of traditional algorithms in terms of data prediction and computational speed, and achieved more efficient energy management and improved utilization of renewable energy.

CN116562423BActive Publication Date: 2026-05-26XI'AN POLYTECHNIC UNIVERSITY
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
XI'AN POLYTECHNIC UNIVERSITY
Filing Date
2023-03-28
Publication Date
2026-05-26

AI Technical Summary

Technical Problem

Existing traditional mathematical programming algorithms and heuristic algorithms suffer from the problem of relying on data prediction accuracy in the optimized operation of electric-thermal coupled new energy systems, and their computational speed is limited, making it difficult to effectively improve the utilization rate of renewable energy.

Method used

By employing a deep reinforcement learning-based approach, an optimized operation model for an electro-thermal coupled new energy system is established. A Markov decision process and reward function mechanism are designed, and an improved multi-threaded PPO algorithm is used to train the agent to optimize the energy management strategy and achieve flexible energy supply for the electro-thermal coupled system.

Benefits of technology

It improves the utilization rate of renewable energy, enhances the energy management efficiency of the system under complex conditions, reduces the reliance on data forecasting, and strengthens the system's peak-shaving capacity and renewable energy absorption capacity.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN116562423B_ABST
    Figure CN116562423B_ABST
Patent Text Reader

Abstract

This invention discloses an energy management method for an electric-thermal coupled renewable energy system based on deep reinforcement learning. The method includes establishing an objective function and response constraints for an optimized operation model of the electric-thermal coupled renewable energy system; expressing the established model as a Markov decision process, defining it as an environment within the deep reinforcement learning framework, and designing a corresponding reward function mechanism; and using an improved multi-threaded PPO algorithm to maximize the cumulative reward to obtain the optimal energy management strategy. This invention's energy management method for an electric-thermal coupled renewable energy system considers pumped storage units and decouples the electrothermal relationship between combined heat and power (CHP) units. During the optimization process, the multi-threaded PPO algorithm based on deep reinforcement learning is applied to the energy management problem of the electric-thermal coupled renewable energy system, which can improve the utilization rate of renewable energy.
Need to check novelty before this filing date? Find Prior Art