Title: Coordinated optimal dispatch of composite energy storage microgrid based on double deep Q-network

Authors: Zheyong Piao; Tianyu Li; Benfa Zhang; Lei Kou

Addresses: Baicheng Power Supply Company, State Grid Jilin Electric Power State, Baicheng, Beijing, China ' School of Electrical Engineering, Changchun Institute of Technology, Changchun, Jilin, China ' State Grid Songyuan Power Supply Company, Songyuan, Jilin, China ' Institute of Oceanographic Instrumentation, Qilu University of Technology (Shandong Academy of Sciences), Qingdao, Jinan, Shandong, China

Abstract: In order to optimise the coordinated control of micro-grid complex energy storage including photovoltaic and wind power, improve the absorption ability of distributed energy generation and reduce the cost, this paper proposes a Double Deep Q-Network reinforcement learning algorithm to train agents to interact with the microgrid environment and learn the optimal scheduling control mechanism. The agent trains the input state and outputs an optimal action to drive the agent's behaviour, including environment perception, action perception and task coordination. It then successfully completes the given task in the complex decision environment. This method can realise multi-objective control for different times, weather conditions and seasons and flexibly process energy storage, hydrogen storage and load energy to achieve coordinated distribution. First of all, a composite energy storage microgrid system model connected to the main power grid is constructed, and deep reinforcement learning activities, state space, reward mechanism and other links are designed. Secondly, in the aspect of learning distributed generation data, a combination of training set and test set of data is proposed for model learning and training. Finally, the optimisation scheduling results of reinforcement learning are analysed for different scenarios of composite energy storage microgrid.

Keywords: microgrid; deep reinforcement learning; composite energy storage; optimise scheduling.

DOI: 10.1504/IJWMC.2024.136576

International Journal of Wireless and Mobile Computing, 2024 Vol.26 No.1, pp.92 - 98

Received: 06 May 2023
Accepted: 27 Jul 2023
Published online: 07 Feb 2024 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article