Training an LSTM Voltage Sags Classificator on a Synthetic Dataset

In this paper, an end-to-end deep learning method for voltage sag classification using a Long Short-Term Memory (LSTM) neural network is proposed. The network receives an unmodified three-phase voltage signal sequence as an input and outputs a label to indicate if a signal is not faulted or there is one of the seven standard voltage sag types present, labeled from A to G respectively. A database o...