摘要
随着新能源发电比例不断提高,电力系统的惯量特性和频率响应能力发生显著变化,快速准确估计系统惯量和一次调频系数对于频率控制和参数优化至关重要。为此提出一种基于变分贝叶斯无迹卡尔曼滤波的电力系统惯量和一次调频系数联合在线估计方法。首先,构建系统频率动态状态空间模型,为后续参数的联合估计奠定基础;然后,综合利用无迹卡尔曼滤波和变分贝叶斯推断实现噪声统计特性的自适应估计;同时引入指数平滑、残差反馈和t-Copula耦合参数更新策略,提高估计的精度和鲁棒性;最后,基于改进IEEE-39节点系统开展算例测试。仿真结果表明所提方法在不同扰动场景下均能快速准确估计系统惯量和一次调频系数,与传统方法相比,在估计精度和鲁棒性方面具有明显优势。
With the continuous increase in the proportion of new energy generation,the inertial characteristics and frequency response capability of power systems have undergone significant changes.Rapid and accurate estimation of system inertia and primary frequency regulation coefficient is crucial for frequency control and parameter optimization.This paper proposes a joint online estimation method for power system inertia and primary frequency regulation coefficient based on the Variational Bayesian Unscented Kalman Filter.First,a dynamic state-space model of system frequency is constructed,which establishes a foundation for the joint estimation of the subsequent parameters.Then,the Unscented Kalman Filter and Variational Bayesian inference are comprehensively utilized to achieve adaptive estimation of noise statistical characteristics.Simultaneously,exponential smoothing,residual feedback,and t-Copula coupling parameter update strategies are introduced to improve the accuracy and robustness of the estimation.Finally,case testing is conducted based on the modified IEEE-39 node system.Simulation results demonstrate that the proposed method can rapidly and accurately estimate system inertia and primary frequency regulation coefficient under various disturbance scenarios,showing significant advantages in estimation accuracy and robustness compared to traditional methods.
作者
欧阳雪彤
陈保瑞
叶希
王曦
杨凯
文云峰
OUYANG Xuetong;CHEN Baorui;YE Xi;WANG Xi;YANG Kai;WEN Yunfeng(State Grid Sichuan Electric Power Company,Chengdu,Sichuan 610041,China;Electric Power Research Institute of State Grid Sichuan Electric Power Company,Chengdu,Sichuan 610047,China;College of Electrical and Information Engineering,Hunan University,Changsha,Hunan 410082,China)
出处
《广东电力》
北大核心
2025年第12期55-68,共14页
Guangdong Electric Power
基金
国家电网有限公司科技项目(52199723003H)。
关键词
惯量
一次调频
变分贝叶斯
无迹卡尔曼滤波
参数估计
inertia
primary frequency regulation
Variational Bayesian
Unscented Kalman Filter
parameter estimation
