| 引用本文: | 朱振山,陈豪,陈哲盛.基于MASAC算法的多园区综合能源系统联合优化调度[J].电力自动化设备,2026,46(2):39-48. |
| ZHU Zhenshan,CHEN Hao,CHEN Zhesheng.Joint optimization and scheduling of multi-park integrated energy system based on MASAC algorithm[J].Electric Power Automation Equipment,2026,46(2):39-48. |
|
| 摘要: |
| 为解决目前需求响应建模不精确及多园区内部市场机制不完善等问题，提出了一种基于多智能体柔性行动器-评判器算法的多园区综合能源系统联合运行优化方法。对各园区内部的热电联供、有机朗肯循环、储能和综合需求响应进行建模；构建多园区系统内部能量市场交易机制，由内部市场管理者根据外部能源价格及多园区系统内能源供需关系确定内部价格；将多园区综合能源系统运行优化问题转化为马尔可夫决策过程，在保证安全运行的前提下以最小化园区自身综合运行成本为目标设计各智能体奖惩函数，采用基于集中训练分散执行框架的多智能体柔性行动器-评判器算法进行求解。仿真验证表明所提联合运行方式及优化算法可以有效降低各园区的运行成本。 |
| 关键词: 综合能源系统 综合需求响应 多智能体深度强化学习 市场机制 运行优化 |
| DOI：10.16081/j.epae.202511020 |
| 分类号: |
| 基金项目:国网福建省电力有限公司科技项目(52130N22000C) |
|
| Joint optimization and scheduling of multi-park integrated energy system based on MASAC algorithm |
|
ZHU Zhenshan1,2, CHEN Hao1, CHEN Zhesheng3
|
|
1.College of Electrical Engineering and Automation, Fuzhou University, Fuzhou 350108, China;2.Fujian Province University Engineering Research Center of Smart Distribution Grid Equipment, Fuzhou 350108, China;3.Fuqing Power Supply Company of State Grid Fujian Electric Power Co.,Ltd.,Fuqing 350300, China
|
| Abstract: |
| To address the issues of inaccurate demand response modeling and imperfect internal market mechanisms of multi-park energy systems, a joint operation optimization method based on a multi-agent soft actor-critic algorithm is proposed. The combined heat and power, organic Rankine cycle, energy storage and integrated demand response of the internal components for each park are modeled. The internal energy market trading mechanism is established, where the internal market manager determines the internal price based on external energy price and the energy supply-demand relationship within the multi-park system. The operation optimization problem of multi-park integrated energy system is formulated as a Markov decision process, under the premise of ensuring safe operation, the reward function for each agent is designed to minimize the comprehensive operating cost of the parks, the multi-agent soft actor-critic algorithm based on a centralized training with decentralized execution framework is adopted to solve the problem. The simulation results verify that the proposed joint operation strategy and optimization algorithm can effectively reduce the operating costs of each park. |
| Key words: integrated energy system comprehensive demand response multi-agent deep reinforcement learning market mechanism operational optimization |
