Deep Reinforcement Learning Based Bi-layer Optimal Scheduling for Microgrids Considering Flexible Load Control  

作　　者：Zitong Zhang Jing Shi Wangwang Yang Zhaofang Song Zexu Chen Dengquan Lin 

机构地区：[1]the State Key Laboratory of Advanced Electromagnetic Engineering and Technology,School of Electronic Engineering,Huazhong University of Science and Technology,Wuhan 430074,China

出　　处：《CSEE Journal of Power and Energy Systems》2023年第3期949-962,共14页中国电机工程学会电力与能源系统学报（英文）

基　　金：supported in part by National Key R&D Program of China under Grant 2021YFB3800200.

摘　　要：In this paper,the bi-layer scheduling method for microgrids,based on deep reinforcement learning,is proposed to achieve economic and environmentally friendly operations.First,considering the uncertainty of renewable energy,the framework of day-ahead and intra-day scheduling is established,and the implementation scheme for both price-based and incentive-based demand response(DR)for the flexible load is determined.Then,comprehensively considering the operating characteristics of the microgrid in the day-ahead and intra-day time scales,a bi-layer scheduling model of the microgrid is established.In terms of algorithms,since day-ahead scheduling has no strict requirement for dispatching time,the particle swarm optimization(PSO)algorithm is used to optimize the time-of-use electricity price and distributed power output for the next day.Considering the environmental fluctuations and requirements for rapidity of intra-day online scheduling,the deep reinforcement learning(DRL)algorithm is adopted for optimization.Finally,based on the data from the actual microgrid,the rationality and effectiveness of the proposed scheduling method is verified.The results show that the proposed bi-layer scheduling based on the PSO and DRL algorithms achieves the optimization of scheduling cost and calculation speed,and is suitable for microgrid online scheduling.

关 键 词：Bi-layer optimal scheduling demand response deep reinforcement learning microgrid scheduling 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TM73[自动化与计算机技术—控制科学与工程]