Fast virtual inertia evaluation for doubly fed induction generator wind farms

Introduction: The high penetration of doubly fed induction generator (DFIG)-based wind farms substantially reduces the equivalent inertia of power systems, rendering the accurate evaluation of their inertia support capability critical for future grid frequency stability. However, continuous wind speed fluctuations pose a major challenge to practical parameter identification. Materials and methods: This paper proposes a fast identification method for DFIG wind farms integrating wind speed fluctuation decoupling and multi-timescale analysis. The proposed scheme consists of three steps: first, an adaptive Savitzky–Golay filter considering the rate of change in frequency performs zero-phase-shift filtering on wide-area measurement data. Second, the inherent frequency regulation characteristics of DFIGs are employed to decouple power response components caused by wind speed variations. Third, multi-timescale decoupling is applied to suppress steady-state drift. Results: The simulation results show that under load disturbance, the proposed method achieves a root-mean-square error of only 0.2017 MW, representing a 55.55% reduction compared with conventional methods. Moreover, its single computation time is merely 8 ms. Conclusions: The results validate that the proposed approach can rapidly and accurately characterize the inertial response of wind farms, offering a reliable and efficient tool for fast inertia assessment to ensure power system stability.