A Non-Iterative Method Based on Fast Fourier Transform and Least Square for Fault Locating in DC Microgrids

Multidimensional Time-Series Shapelet-Based Real-Time Fault Detection and Localization on ISS Electrical Power Distribution System

“…Timeseries features are obtained by simply normalizing without complex computations [14], but redundant and disturbing information contained in these features may increase difficulties of learning principles by the FDL method. Time-domain and frequency-domain features are commonly extracted from raw time series using signal processing methods, such as wavelet transform [15], fast Fourier transform (FFT) [16], Hilbert-Huang transform (HHT) [17] and so on. However, signal processing is usually time-consuming and elaborate feature selection is required.…”