| 引用本文: | 孙正龙,李慎浩,李泽伟,等.含风电电力系统的智能广域阻尼控制策略研究[J].电力系统保护与控制,2026,54(05):144-154. |
| SUN Zhenglong,LI Shenhao,LI Zewei,et al.Intelligent wide-area damping control strategy for wind-integrated power systems[J].Power System Protection and Control,2026,54(05):144-154 |
|
| 摘要: |
| 高比例风电接入加剧了风电机组与同步机间的动态耦合，导致多机广域阻尼控制器(wide area damping controllers, WADC)难以协同整定。在复杂工况下，多机WADC难以有效抑制振荡，甚至恶化系统阻尼。为解决上述问题，提出一种面向含风电电力系统的智能广域阻尼控制策略。首先，建立含双馈风机的多机广域阻尼协同控制模型，并基于联合测度指标选择广域阻尼控制回路。然后，构建融合主成分分析(principal component analysis, PCA)与多智能体深度确定性策略梯度(multi agent deep deterministic policy gradient, MADDPG)的控制框架，实现高维状态空间下的参数协同优化。最后，在改进的两区域四机系统及我国西北风火打捆送端系统开展仿真分析，结果表明该策略能够显著提升区域间低频振荡模态的阻尼水平，在复杂工况下依然具备良好的振荡抑制效果和工程适应性。 |
| 关键词: 高渗透率风电 广域阻尼控制 MADDPG 多智能体强化学习 控制器参数协同整定 主成分分析 |
| DOI：10.19783/j.cnki.pspc.250778 |
| 分类号: |
| 基金项目:智能电网国家科技重大专项(2030)资助(2024 ZD0801000)；国网浙江省电力有限公司科技项目资助(B311DS25000A) |
|
| Intelligent wide-area damping control strategy for wind-integrated power systems |
|
SUN Zhenglong1, LI Shenhao1, LI Zewei1, ZHANG Rui1, YANG Hao1, HUA Wen2, ZHANG Chengming2
|
|
1. Key Laboratory of Modern Power System Simulation and Control & Renewable Energy Technology (Northeast
Electric Power University), Ministry of Education, Jilin 132012, China; 2. State Grid Zhejiang Electric
Power Research Institute, Hangzhou 310014, China
|
| Abstract: |
| High wind power penetration increases dynamic coupling between wind turbine generators and synchronous machines, making coordinated tuning of wide-area damping controllers (WADCs) challenging. Under complex operating conditions, multiple WADCs may fail to suppress oscillations and may even deteriorate system damping. To address these issues, an intelligent WADC strategy for wind-integrated power systems is proposed. First, a coordinated control model with multiple doubly-fed induction generators (DFIGs) is established, and control loops are selected using a composite geometric index. A hybrid control framework is then developed by integrating principal component analysis (PCA) with the multi-agent deep deterministic policy gradient (MADDPG) algorithm to optimize parameters in high-dimensional state spaces. Finally, simulations on a modified two-area four-machine system and a wind-thermal sending-end system in Northwest China show that the proposed method significantly improves inter-area low frequency damping and maintains effective oscillation suppression performance and engineering adaptability under complex conditions. |
| Key words: high wind power penetration wide-area damping control MADDPG multi-agent reinforcement learning coordinated controller parameter tuning principal component analysis (PCA) |
