Key points are not available for this paper at this time.
Online identification of the stability region for the wind power plant (WPP) equipped with permanent magnet synchronous generators encounters two challenges: calculating high-order impedance and estimating the stability boundary of multiple parameters. To address these challenges, the mechanism-data-driven equivalent impedance for each branch of the WPP is developed, and the high-dimensional stability region boundary is established with a machine learning-based algorithm. In mechanism-driven modeling, the aggregated impedance is derived using parallel calculation. In data-driven modeling, different key parameters in wide frequency bands are modified to ensure the accuracy of the equivalent impedance. The equivalent impedance derived through the mechanism-data-driven approach deviates from the detailed impedance within 1\%. Regarding the stability region estimation, the fuzzy support vector machine is applied. To improve the speed of the online training of the stability region, the weighted Mahalanobis distance is employed to distinguish the importance of stability index samples. In validation, the stability regions of the power of the WPP are established. The impacts of the static var generator (SVG) on the stability performance are analyzed, and regulating the reactive power of the SVG can enhance small-signal stability. The stability region estimation and enhancement are verified through time-domain simulations and experiments in a controller-hardware-in-loop platform.
Qin et al. (Mon,) studied this question.