基于数据驱动的双馈风电机组阻抗辨识及稳定性分析

Data-Driven Impedance Identification and Stability Analysis of Doubly-Fed Induction Generator Based Wind Turbine

阻抗分析法是研究“双高”电力系统宽频振荡问题的有效工具,而传统阻抗建模方法存在需要提前获取系统详细结构及控制参数、建模过程复杂、不便于在线应用等不足。为解决上述问题,提出了一种基于改进BP神经网络的双馈风电机组阻抗模型辨识方法。定义了评估系数,以评估神经网络辨识模型的准确度,并基于评估系数改进了BP神经网络的结构,在较优的网络结构下,可用更少的数据实现准确的多工况阻抗模型辨识。通过对比所提方法与传统BP神经网络方法的辨识阻抗误差,说明了所提方法的优越性。为验证所提方法在稳定性在线分析中的可行性,基于辨识阻抗分析了不同电网短路比和不同风电机组出力下双馈风电机组并网系统的稳定性,时域仿真结果验证了所提方法的准确性。

The impedance-based analysis method is an effective tool for studying the wideband oscillation problems of“double-high”power systems.However,traditional impedance modeling methods have deficiencies such as the necessity to obtain a detailed structure and control parameters of the system in advance,complex modeling process,and inconvenience for online applications.To address these challenges,an improved BP neural network based impedance model identification method for a doubly-fed induction generator(DFIG)based wind turbine is proposed herein.An evaluation coefficient was defined to evaluate the accuracy of the neural network identification model,and the structure of the BP neural network was improved based on this.Under the improved network structure,accurate multi-condition impedance model identification can be achieved with less data.By comparing the identification impedance errors of both the proposed and the traditional BP neural network methods,the superiority of the proposed method was demonstrated.To verify the feasibility of the proposed method for online stability analysis,the stability of the DFIG-based wind turbine grid integration system under different grid short-circuit ratios and wind turbine output active power conditions was analyzed using the identified impedance.