An intelligent machine learning-based algorithm to adapt behavior of differential protection during sub synchronous resonance in smart grids

The smart grid incorporates renewable energy generation sources, such as Doubly Fed Induction Generators (DFIG) in wind farms, and emphasizes the importance of self-healing capabilities, which enhance the grid’s reliability. Sub-Synchronous Resonance (SSR) caused by series compensation is a significant concern in transmission lines, as it leads to substantial increases in voltage and current magnitudes. Additionally, SSR raises the likelihood of ferroresonance occurring within the grid. This phenomenon can disrupt the differential protection systems of transformers, busbars, and short lines in transmission substations. The latest intelligent protective algorithms proposed in the literature can recognize SSR and adapt various types of relays during SSR. However, these algorithms have not yet been thoroughly investigated in our study. The smart grid requires self-healing protection; therefore, this paper proposes an intelligent machine learning-based algorithm capable of adapting the behavior of transformer differential protection during SSR. The algorithm utilizes Discrete Wavelet Transform (DWT) and a variety of machine learning techniques to distinguish SSR and ferroresonance caused by SSR from other similar phenomena. It determines whether to operate or restrict the relay by modifying setting parameters according to the protection strategy, which can also be derived using intelligent techniques.