With the increasing penetration of wind power into modern power systems, wind energy-integrated systems exhibit pronounced probabilistic uncertainty under the combined effects of stochastic wind speed fluctuations and small-signal disturbance. Traditional small-signal stability analysis methods based on deterministic models are therefore insufficient to support real-time control and stability assurance of wind turbines. To address the limitations of conventional offline small-signal stability assessment in providing effective real-time control decisions under high wind power penetration, an integrated “analysis-decision-control” framework oriented toward controllable probabilistic stability was proposed. First, leveraging observations from PMUs (phasor measurement units), AKDE (adaptive kernel density estimation) approach was employed to construct probabilistic small-signal models of key system states—such as frequency, voltage, and current—under wind power integration, enabling a refined characterization of non-Gaussian disturbance features in stochastic environments. Second, based on the results of probabilistic stability analysis, a Markov decision process was formulated with the output power as the system state and pitch angle regulation as the control action. ABERL (adaptive bayesian ensemble reinforcement learning) control strategy was then developed, in which multiple policies were adaptively fused through posterior inference to achieve self-adaptive pitch angle control of wind turbines, thereby directly optimizing the probabilistic small-signal stability of the system. Finally, the proposed method was validated on the IEEE 10-machine and 39-node typical system and the real-world Southern Power Grid system under PMU-driven small-disturbance scenarios, and was compared with representative control strategies including Actor-Critic, Monte Carlo methods and Q-learning. Simulation results and case study demonstrate that the proposed approach achieves superior performance in terms of disturbance response speed, pitch angle control stability, and probabilistic stability metrics. The system-level probabilistic stability analysis effectively identifies potential instability risks, while precise pitch angle regulation of local wind turbines significantly reduces wind farm control costs and enhances stable operation capability. The proposed probabilistic-driven control framework thus provides a viable solution for the secure and stable operation of power systems with high wind power penetration.
YU Miao, YAN Jia-xin, HAN Jin-yang, WU Hong-hao, ZENG Run-xin.
Probabilistic Stability of Small-signal Disturbance in Wind Power System Based on Adaptive Bayesian Ensemble Reinforcement Learning[J].
Science Technology and Engineering, 2026, 26(18): 7730-7740 DOI:10.12404/j.issn.1671-1815.2503956
在小信号稳定性分析领域,概率方法因能够将系统输入随机变化纳入考虑来建模不确定性,并评估不同结果的概率,表现出显著优势。对动态稳定性概率方法最先通过近似多元正态分布进行探索,该方法结合特征值敏感性来计算特征值统计量。相比MC方法,这种基于特征值的概率方法在精度和效率上均有优势。在不同风电渗透率水平下,线性化特征值敏感性与风电概率特性相结合,利用最大熵理论研究概率小信号稳定性,在精度和准确性上均有所提高[17]。为减少MC方法所需大量计算,文献[18]提出概率方法可以直接计算关键特征值的概率密度函数(probabilistic density function, PDF)以评估风电并网对电力系统小信号稳定性的影响。假设风速服从Weibull分布前提下,一种小信号稳定性概率方法通过Gram-Charlier展开推导出关键系统特征值PDF,对Weibull分布的依赖性限制其适用性。文献[19]提出一种基于矩阵铅笔方法新框架,用于分析各种数值方法在小信号稳定性分析中误差与稳定性。通过比较离散时间系统模态与实际动态系统模态,该框架提供对小信号扰动条件下数值方法在时域仿真中表现差异的更深入理解。此外,文献[20]通过不同控制策略进行频率概率稳定性分析,并在风电动态过程中进行验证。在概率分析中提供不确定性定量评估后,适当的控制措施可以迅速调整关键系统参数,确保随机环境稳定性[21]。
由于PMU设备以高频率采集电压、电流和相位等连续数据,其数据常含噪声与突变,传统方法难以有效建模其复杂动态特性。为准确地刻画风电数据中因波动性引起的局部密度差异,引入AKDE方法。通过在数据密集区采用较小带宽以捕捉细节结构,在数据稀疏区使用较大带宽以增强整体平滑性,自适应带宽 KDE 能有效提升局部特征敏感性。
(1)针对随机环境下风电系统小扰动响应的概率不确定性刻画,提出一种基于PMU观测数据的自适应核密度估计(adaptive kernel density estimation, AKDE)概率建模方法,对风电并网条件下频率、电压、电流等关键状态量进行非参数概率建模。通过引入自适应带宽机制,有效提升非高斯扰动及状态分布非均匀区域建模精度,为后续概率稳定性分析与控制决策提供可靠状态描述基础。
WangQuan, ZhangZhaogang, XuTangjie, et al. Synchronous control optimization of wind turbine start-stop and yaw in wind farm[J]. Science Technology and Engineering, 2024, 24(34): 14666-14671.
HuQiguo, WangZelin, HuHuoran. Application of active disturba-nce rejection control based pitch system optimization in wind power system[J]. Science Technology and Engineering, 2023, 23(28): 11985-11992.
[5]
YadavV V, SaravananB. Technical advances and stability analysis in wind-penetrated power generation systems: a review[J]. Frontiers in Energy Research, 2022, 10: 1091512.
TianJingjing, XingKang. Vulnerability analysis of wind power and photovoltaic integration based on complex network and electrical characteristics[J]. Science Technology and Engineering, 2025, 25(6): 2379-2388.
[8]
HuZ X, XuY U, KorkaliM, et al. A Bayesian approach for estimating uncertainty in stochastic economic dispatch considering wind power penetration[J]. IEEE Transactions on Sustainable Energy, 2021, 12(1): 671-681.
[9]
LiuY L, WangJ Y, YueZ Y. Improved multi-point estimation method based probabilistic transient stability assessment for power system with wind power[J]. International Journal of Electrical Power & Energy Systems, 2022, 142.DOI: 10.1016/j.ijepes.2022.108283.
[10]
ZhangT, XiaoS P. Stability analysis of the sampled-data control system and its application to the electric power market[J]. Chinese Journal of Electrical Engineering, 2020, 6(1): 61-70.
YuMiao, HuJingxuan, ZhangShouzhi. Fast stability of new power system based on a PMU gradient dynamic deviation method[J]. Journal of Shanghai Jiao Tong University, 2024, 58(1): 40-49.
[13]
LiL M, JiangH, YangP H, et al. Small signal stability analysis and optimize control of large-scale wind power collection system[J]. IEEE Access, 2022, 10: 28842-28852.
[14]
GustavoF G, MohammadK B, et al. State space modeling of an offshore wind power plant with an MMC-HVDC connection for an eigenvalue-based stability analysis[J]. IEEE Access, 2022, 10: 82844-82869.
[15]
YuM, LiJ L, ZhangS Z. Probabilistic stability of small disturbance in wind power system based on a variational Bayes and Lyapunov theory using PMU data[J]. IET Generation, Transmission & Distribution, 2022, 16(23): 4818-4829.
[16]
NkosiN R, BansalR C, AdefaratiT, et al. A review of small-signal stability analysis of DFIG-based wind power system[J]. International Journal of Modelling and Simulation, 2022, 43(3): 153-170.
[17]
PepicielloA, GarciaJ D. Small-signal stability analysis of uncertain power systems: a comprehensive survey[J]. Renewable and Sustainable Energy Reviews, 2024, 200: 114576.
GongHaoyue, GuoQiang, GuoJianbo, et al. Research on small-signal stability analysis and control based on converter system state-space method[J]. Proceedings of the CSEE, 2025, 45(13): 4943-4955.
[20]
YangL, ZhangH G, GaoZ Y. Application of small disturbance instability based on random response surface method in wind power system[C]// 40th Chinese Control Conference. Shanghai: CCC, 2021: 2487-2492.
[21]
DingY, YeX W, GuoY, et al. Probabilistic method for wind speed prediction and statistics distribution inference based on SHM data-driven[J]. Probabilistic Engineering Mechanics, 2023, 73: 103475.
[22]
CaoR L, XingJ, LiZ, et al. Probabilistic small signal stability analysis with wind power based on maximum entropy theory[J]. Electric Power Components and Systems, 2024.DOI: 10.1080/15325008.2024.2310775.
[23]
WangZ Y, BuS Q. Probabilistic analysis of small-signal stability of power systems affected by wind power generation uncertainties considering wake effects[J]. IEEE Transactions on Power Systems, 2025, 40(2): 1955-1968.
LiLixin, TianXu, LiuFei, et al. Analysis ofwind power fluctuation characteristics based on hybrid adaptive kernel density estimation[J]. Renewable Energy, 2024, 42(7): 923-928.
[26]
WangZ Y, BuS Q. Probabilistic frequency stability analysis considering dynamics of wind power generation with different control strategies[J]. IEEE Transactions on Power Systems, 2024, 39(5): 6412-6425.
[27]
DongH Y, ZhangJ C, ZhaoX W. Intelligent wind farm control via deep reinforcement learning and high-fidelity simulations[J]. Applied Energy, 2021, 292: 116928.
HeTing, QiaoJunqiang, WuGuodong. Stability analysis and parameter optimization of virtual synchronous generator based on small-signal model[J]. Science Technology and Engineering, 2025, 25(5): 1927-1935.
[30]
FangJ H, HuW F, LiuZ Y, et al. Wind turbine rotor speed design optimization considering rain erosion based on deep reinforcement learning[J]. Renewable and Sustainable Energy Reviews, 2022, 168: 112788.
[31]
JesusE, MatildeS. Exploring reward strategies for wind turbine pitch control by reinforcement learning[J]. Applied Sciences, 2020, 10(21): 7462.
[32]
XieJ J, DongH Y, ZhaoX W. Data-driven torque and pitch control of wind turbines via reinforcement learning[J]. Renewable Energy, 2023, 215: 118893.
[33]
GuanJ S, LinJ, GuanJ J, et al. A novel probabilistic short-term wind energy forecasting model based on an improved kernel density estimation[J]. International Journal of Hydrogen Energy, 2020, 45(43): 23791-23808.
[34]
ZhongT W, ZhangH T, LiY Z, et al. Bayesian learning-based multi-objective distribution power network reconfiguration[J]. IEEE Transactions on Smart Grid, 2021, 12(2): 1174-1184.
[35]
JasminB. A review of Bayesian machine learning principles, methods, and applications[J]. International Journal of Innovative Science and Research Technology, 2023, 8(5): 2033-2038.
RenJun, YuePengfei, ZhangZhiyu, et al. Analysis based on the data preprocessing method of wind turbines based on isolated forest method[J]. Science Technology and Engineering, 2024, 24(6): 2378-2384.
WuJing, ChengWenjuan, LiangXiao, et al. Game and strategy analysis of power price incentive response frequency modulation in load aggregation based on reinforcement learning[J]. Science Technology and Engineering, 2025, 25(3): 1087-1092.