Electric ship digital twin : framework for cyber-physical system security

Abstract

This thesis presents a dynamic model that can be used for a digital twin of an electric ship. The model is an end-to-end simulation of a ship from prime mover to maneuvering, seakeeping, and propeller ventilation in random waves representing the behavior of a physical ship. There is a trend towards increasingly networked sensors and actuators to enable condition monitoring, ensure efficient operation, and allow for autonomy. However, a cyberattack on a networked control system presents not just the possibility of information theft but of physical system damage and loss of control. Thus, a detection scheme is proposed for cyber-physical systems using a joint unscented Kalman filter. It is employed to detect cyberattacks in the simulation model of an electric ship including sensor attacks and controller attacks on a gas turbine, synchronous generator, and automatic heading control. Finally, a system theoretic framework is presented for optimal sensor placement to minimize cyber vulnerability.