An Identification Method for Sequence Impedance Model of DFIG Based on the Joint of Knowledge Driving and Data Driving

Abstract

Impedance based stability analysis is an effective tool to study broadband oscillation in power system. However, the analytical modeling process of impedance of renewable energy equipment is complex, and the impedance acquisition method based on data measurement has the defects of time-consuming and inconvenience for online measurement. In order to solve the above problems, an impedance identification method based on the joint of knowledge driving and data driving for Doubly-Fed Induction Generator (DFIG) is proposed. First, the small-signal model of DFIG is established based on knowledge-driven, acquiring the variables with nonlinear relationship with sequence impedance. Then, the Extreme Gradient Boosting (XGBoost) model is trained based on data driving to realize the identification of DFIG's sequence impedance under multiple operating conditions. Finally, the experiment on CHIL (Control-hardware-in-loop) is carried out to verify the accuracy of the XGBoost model.