基于深度强化学习的微电网优化运行策略. (Chinese)

Optimal operation strategy of microgrid based on deep reinforcement learning. (English)

The uncertainty of wind power，photovoltaic and load brings challenges to the formulation of opera‐ tion strategy for microgrid with high proportion of renewable energy，and the development of artificial intelli‐ gence technology provides a new idea for solving the operation optimization problem of microgrid. Based on the reinforcement learning framework，the operation problem of microgrid is transformed into a Markov decision process，and an online scheduling method of microgrid based on deep reinforcement learning is proposed，which takes the maximum economic benefit of microgrid and residents’satisfaction as its object. In order to effectively use the experience in the training process of deep reinforcement learning，a PESDDPG（Priority Experience Storage Deep Deterministic Policy Gradient） algorithm is designed to learn the optimal scheduling strategy of microgrid for different periods under each type of environment. Case results show that PES-DDPG algorithm can provide effective scheduling strategy for microgrid and realize real-time optimization of microgrid. [ABSTRACT FROM AUTHOR]

风电、光伏、负荷的不确定性给含有高比例可再生能源的微电网制定运行策略带来了挑战，人工智能技 术的发展为解决微电网运行优化问题提供了新思路。基于强化学习框架，将微电网运行问题转化为马尔可 夫决策过程，以最大化微电网经济利益和居民满意度为目标，提出一种基于深度强化学习的微电网在线调度 方法。为了在深度强化学习训练的过程中高效利用经验，设计一种优先经验存储的深度确定性策略梯度 （PES-DDPG）算法，学习各类环境下不同时段的微电网最优调度策略。算例结果表明，PES-DDPG 算法能够 为微电网提供有效的调度策略，并实现微电网的实时优化. [ABSTRACT FROM AUTHOR]

Copyright of Electric Power Automation Equipment / Dianli Zidonghua Shebei is the property of Electric Power Automation Equipment Press and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)