Integrated Approach for Model-based Adaptation and Data-drive Learning for Ship Power Network Health Monitoring and Prognosis
Abstract
Maintaining continuous operation of power and energy systems of Navy ships at a high performance level is of paramount priority for military force readiness and mission effectiveness. It requires an effective condition monitoring system to detect deterioration, an intelligent adaptive system to compensate anydeterioration if detected, and a prognostic system to make decisions on proactive actions to prevent downtime and minimize cost. For all-electric ships, the increased system complexity due to electrification and the interactive nature of energy subsystems associated with integrated power systems (IPS) have led to a completely new set of requirements for shipboard condition monitoring, particularly for those with substantial onboard energy storage devices. The project proposed herein is in response to the new challenges for maintaining continuous operation of all-electric ships. It aims to develop a comprehensive analysis and design framework for shipboard power network condition monitoring and prognosis. Built on results from prior ONR supported projects, and leveraging our expertise on modeling, system identification, control, and adaptation, the team will (a) develop methodologies and tools to enable accurate and on-time detection of performance deterioration of shipboard power networks; (b) develop analytical concepts, numerical algorithms, and experimental evidence for active condition monitoring and prognosis of shipboard power networks; (c) develop design guidelines for control architecture, learning mechanisms, and adaptation strategies for highly interactive power and energy systems. Focusing on the characterization of SoX (namely, the state of charge (SoC) for energy storage, the state of health(SoH) for power and energy components, and the state of power capability (SoP)), we will explore advanced model-based identification and data-driven learning mechanisms to achieve real-time SoX estimation at both the component and system levels. The developed concepts and algorithmswill be demonstrated using the Advanced Electric Drive and Hybrid Energy Storage (AEDHES) testbed at the University of Michigan. Results of this work will be shared through close collaboration with the Center for Advanced Power Systems (CAPS) at Florida State University, where the developed techniques will be applied at a ship-level scale.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jul 10, 2018
- Source ID
- N000141812330
Entities
People
- Jing Sun
Organizations
- Board of Regents of the University of Michigan
- Office of Naval Research
- United States Navy