以往在对双馈风电机组进行同调判别时,常将发电机端口的有功功率及无功功率受扰轨线作为观测量。但由于有功功率及无功功率不是系统的状态量,将其作为同调判别的观测量可能带来一定误差。基于同步发电机及异步电动机的同调判别方法,指出应选择主导模式中参与因子较大的状态量作为观测量。即基于特征根技术,提出双馈风电机组的同调判别应以转子电流为观测量。同时,还提出综合风电机组初始运行点及动态特性的两步分群方法,即先根据风电机组的初始风速进行首次分群,然后基于风电机组的动态特性再次分群。该方法不仅能够较准确地估算等值机的等效风速,还可以使等值系统的动态特性与原系统尽可能一致。最后通过仿真算例验证了所提方法的可行性。
In the existing research on coherence evaluation of doubly fed induction generators (DFIG), the active power and reactive power disturbed trajectories are often used as measurements. However, it may cause error since active power and reactive power are not state variables of the system. Based on coherence evaluation methods for synchronous generators and asynchronous motors, it is pointed out that the state variables with larger participation factors in the dominant mode should be taken as measurements. Based on the eigenvalue technique, it is proposed that the rotor current is suitable for coherency evaluation of DFIG. Meanwhile, a two-step clustering method is proposed. The first step clusters the DFIGs according to their initial wind speed, and the second step re-clusters the DFIGs by their dynamic characteristics. Using this method, not only the equivalent wind speed of the aggregated DFIG can be estimated accurately, but also the dynamic characteristics of the aggregated DFIG are consistent with original system. The proposed clustering method is validated by simulation results.