首页|基于可学习模型预测控制的含风电多微网频率控制方法

基于可学习模型预测控制的含风电多微网频率控制方法

扫码查看
负荷波动、分布式电源出力的强随机性以及多微电网更为复杂的结构设计给微电网负荷频率稳定控制带来了巨大挑战.因此,提出了基于可学习模型预测控制的含变速恒频双馈风机多微电网负荷频率控制(LFC)方法.首先,搭建了包含风力发电机组、储能、微型燃气轮机及负荷的互联多微电网负荷频率控制模型.其次,提出了一种可学习的模型预测控制算法,其能够基于深度强化学习对模型预测控制器实现参数自适应.最后,通过仿真验证了所体控制方法具备在线学习及经验回放能力,在不同的复杂运行工况下均能实现频率的稳定控制.
Load Frequency Control Method for Multi Microgrid with Wind Power Based on Learnable Model Predictive Control
The strong randomness of load fluctuations and distributed power generation as well as the more complex structural design of multiple microgrids pose significant challenges to the load frequency stability control of microgrids.Therefore,the paper propose a load frequency control(LFC)method for the multiple microgrids with variable-speed constant-frequency doubly-fed wind turbines based on learnable model predictive control.Firstly,an interconnected multiple microgrid LFC model that includes wind power generation units,energy storage,micro gas turbines,and loads is constructed.Then a learnable model predictive control algorithm is proposed,which can achieve parameter adaptation for the model predictive controller based on deep reinforcement learning.Finally,some simulations are conducted to verify that the control method has online learning ability and experience replay function,and can achieve stable frequency control under various complex operating conditions.

multi microgrid systemLFCdeep reinforcement learningmodel predictive control

张磊光、陈海涛、吴赋章、杨军

展开 >

国网驻马店供电公司,河南驻马店 463000

武汉大学电气与自动化学院,湖北武汉 430010

多微电网系统 负荷频率控制 深度强化学习 模型预测控制

国家重点研发计划资助项目

2023YFB2407300

2024

智慧电力
陕西省电力公司

智慧电力

CSTPCD北大核心
影响因子:0.831
ISSN:1673-7598
年,卷(期):2024.52(10)
  • 15