Enhanced fault classification in inverter-fed transmission lines using deep learning

Abstract The conventional power transmission system with synchronous generators is protected with phase angle-based approach that decides which type of faut is occurred on the transmission line system. However, the phase angle-based fault classification method is no longer applicable for systems that have integrated Inverter Based Resources (IBR). The control technology employed for grid integration of renewable sources controls the fault current magnitude depending upon Grid code requirements which results in mal operation of relay. A new fault classification scheme is proposed in this paper which aims at precise faulty phase selection using deep learning techniques. Deep learning techniques have gained significance in the field of protection because of the big data availability from Phasor Measurement Units (PMU’s). This paper describes a new approach of fault classification for systems connected to renewables with the application of Long Short-Term Memory (LSTM) Deep learning network.