Early warning of low-frequency oscillations in power system with Vinnicombe criterion fused with PMU data

With the development of the electricity industry in China, warning of low-frequency oscillations has become an important topic in power system stability research. In this paper, we propose a low-frequency oscillation warning method based on the downscaled Vinnicombe distance of key feature data. The method first screens and reduces the dimensionality of the PMU signals collected from the power system, initial feature matrix, and multi-feature vector state detection matrix for low-frequency oscillation amplitude warning indicators generated by key feature data extraction. The transfer function distance is calculated by computing the Vinnicombe distance, which judges whether low-frequency oscillation occurs and effectively improves the identification accuracy. Finally, algorithmic simulation through the New England 10-machine 39-node system is compared with the LOF method.