基于RL-MPC算法的含风电互联电网负荷频率控制研究

Research on load frequency control of interconnected power grid with wind power based on RL-MPC algorithm

摘要

为解决高比例新能源并网引起的电力系统的频率波动问题，提出一种基于强化学习-模型预测控制(RL-MPC)算法的含风电互联电网负荷频率控制策略。首先，考虑风速出力变化引起的模型动态参数，构建了风电参与的两区域负荷频率控制模型；然后，采用强化学习实时优化系统中的模型预测控制的参数，减少控制方案对准确模型的依赖，保证在风电输出不确定的情况下，控制器能够针对系统参数变化进行自适应调整，以保证最优的频率控制效果；最后，通过算法仿真对比验证了所提控制策略的有效性。

Abstract

To address the frequency fluctuations in the power system caused by a high proportion of new energy integration, this paper proposes a load frequency control of interconnected power grid with wind power based on reinforcement learning model predictive control(RL-MPC) algorithm. Firstly, considering the dynamic parameters of the model caused by changes in wind speed output, a two region load frequency control model with wind power participation was constructed. Then, reinforcement learning is used to optimize the parameters of model predictive control in the system in real time, reducing the dependence of control schemes on accurate models and ensuring that the controller can adaptively adjust to changes in system parameters in the case of uncertain wind power output, to ensure the optimal frequency control effect. Finally, the effectiveness of the proposed control strategy was verified through algorithm simulation comparison.

关键词

负荷频率控制 / 风电不确定性 / 模型预测控制 / 强化学习

Key words

load frequency control / wind power uncertainty / model predictive control / reinforcement learning

Author summay

吴向权(1977—), 男, 高级工程师, 硕士, 主要从事电力系统工程项目。E-mail: [e-mail]

基金资助

云南省红河州永宁风电场项目综合自动化系统设备采购项目