A Decentralized Droop-Based Power Management System for Ship Power Systems Using Hybrid Dynamical Systems Framework

Abstract

In maritime transport, inspired by the automobile sector, autonomy is gaining traction in parallel with decarbonization. One of the numerous challenges in realizing fully autonomous operation in shipping is to design a resilient and fault-tolerant power system that preserves the survivability of ships during worst-case failures in unpredictable maritime weather conditions. In newly built ships, power systems are designed with a high number of sensors and communication equipment to enable remote control and condition monitoring in real time. In such power systems, the traditional concept of a centralized power management system (PMS) is not reliable during communication failures and cyberattacks. To address this issue a decentralized fault-tolerant droop-based PMS that does not rely on communication between energy sources is proposed. The droop curves are further designed for the derating operation of energy sources and energy storage devices. A ship power system exposed to faults represents a hybrid system that consists of interaction between continuous and discrete states. Hybrid dynamical systems theory is used to model the DC power system and implement the proposed PMS. The normal operation of energy sources, energy storage devices, and shiploads are modeled as continuous dynamics. The faults such as derating operation and disconnection of energy sources, energy storage devices, and shiploads are modeled as discrete events. The results demonstrate that the proposed PMS can keep the system parameters such as DC bus voltage within the limits permissible by class rules during the loss of power generation.