Classification of Faults on the Shipboard Distribution Power System

This paper discusses the issue of electrical power quality and how it affects the ship safety. Since the new techniques for producing and utilizing electrical energy in the ship systems have been introduced, there is a need to consider the increase in the significance of Power Quality (PQ). The objective of this paper is to classify the PQ disturbances on the ship using Ensemble Bagged Tree, Nearest Neighbors and Support Vector Machine. Hence, the electrical model on the ship distribution system is develop based on the real measurement data. The types of PQ disturbances that has been considered in this classification scheme are voltage sag, voltage swell, and combination of voltage swell and voltage transient. In order to get the high performances in the classification scheme, the S-Transform (ST) is chosen to extract the significant features used by the classifiers. In this case, 60% of total data is used for training and the remaining data will be used for testing. The results shows that the Ensemble Bagged Tree presents high accuracy rate of 91.7% compared to the Nearest Neighbors and Support Vector Machine.