独立变桨距风力发电机系统辨识研究

SYSTEM IDENTIFICATION OF INDIVIDUAL PITCH CONTROL WIND TURBINE

主要研究了独立变桨距风力发电机的系统辨识技术。针对风力发电机系统的非线性,以风速为依据将运行过程分为若干工作点,对每个工作点的线性模型进行辨识;针对独立变桨距控制输入输出的周期性变化,将节距角信号分解为平均分量和差动分量,并将3个周期变化的差动分量通过d-q坐标变换转换为时不变的d轴分量和q轴分量;针对辨识过程中的有色噪音,采用增广的最小二乘法(RELS)对系统进行无偏一致的估计。辨识实验的对象采用成熟的非线性风力发电机模型FAST,在Matlab/Simulink中搭建了辨识实验系统。模型验证结果显示,辨识模型在工作点能很好地拟合原系统的特性,同时也可看出,辨识模型实现了对输入输出的自动解耦,可方便地进行变桨距控制器设计。

The system identification of the individual pitch control wind turbine was presented. Because of the high nonlinearity of wind turbine, a series of models at some stable operating points were identified based on the wind speed. The pitch angles of the stable operating points was periodic value varying with the azimuth, so the pitch angles were divided into one average value and three different values, and d-q transform was utilized to transfer the three different values to two time-invariant values： d axis component and q axis component. RELS method was used as the identification method. The nonlinear wind turbine model FAST was adopted as the identified plant, and Matlab/Simulink was used as the identification tool. The results showed that the identified models can describe the nonlinear models very well at particular operating points; the identified model automatically decoupled input and output, and can be used for independent pitch control design conveniently.