Wavelet based method for fault detection in Medium Voltage DC shipboard power systems

This paper proposes a wavelet transform (WT) based multi-resolution analysis (MRA) method to obtain the features of different fault types in Medium Voltage DC (MVDC) shipboard power systems. DC topology is under consideration for future all-electric ships. One of the new challenges related to this architecture is fault detection. WT-based MRA method, as well as its properties, are studied and adopted for fault detection in this paper. The energy of the wavelet coefficients is chosen to generate the feature vectors. The Daubechies 10 (db10) wavelet and scale 9 are the chosen appropriate wavelet function and optimal decomposition level. Short circuit faults on both DC bus and AC side are studied to verify the proposed method. These faults are simulated in RTDS with a notional MVDC shipboard power system model and the obtained data is then analyzed in MATLAB. The simulation results indicate that promising feature vectors can be extracted to distinguish different faults that may occur on board of a ship.