Machine learning techniques for diagnosing and locating faults through the automated monitoring of power electronic components in shipboard power systems

Abstract

The management and control of shipboard medium voltage AC (MVAC) and medium voltage DC (MVDC) power system architectures under fault conditions present a number of challenges. The use and resulting interaction of multiple power electronic components in mesh-like power distribution architectures possibly result in the effects of faults being detectable throughout the system, for example, line-to-hull faults on DC systems with high resistive grounding.