Dynamic monitoring and identification method for sub-synchronous oscillation in wind power Grid-Integrated systems under a Three-Level collaborative architecture

The frequency of sub-synchronous oscillation (SSO) incidents has increased due to the ongoing expansion of wind power integration into power systems, which presents substantial challenges to the secure and reliable operation of power grids. Conventional methods often lack sufficient resistance to noise and fail to achieve high recognition precision. This study proposes a dynamic SSO monitoring technique that employs a three-tier collaborative architecture for wind power grid-connected systems. The architecture comprises a Low-Computational Monitoring tier, an Adaptive Decomposition tier, and a High-Resolution Identification tier, designed to address these limitations. The Low-Computational Monitoring tier utilizes the periodic oscillation characteristics of three-phase instantaneous power for dynamic oscillation detection. If an SSO is detected, it triggers the subsequent tiers; otherwise, monitoring continues. The Adaptive Decomposition tier employs an enhanced variational mode decomposition (VMD) algorithm to dynamically determine the optimal decomposition parameters, reduce noise, and perform adaptive signal decomposition. The High-Resolution Identification tier applies the Hilbert transform and nonlinear least squares method to the intrinsic mode functions (IMFs) derived from the previous tier to extract instantaneous parameters, thereby enabling precise SSO parameter identification within this collaborative framework. Simulation results demonstrate that the proposed method effectively mitigates mode mixing, exhibits substantial noise immunity, and achieves excellent identification accuracy and robustness.