Resilient Microgrids on Naval Installations by Adoption of Shipboard Zonal Distribution Philosophy Through the Use of Digital Twins
Abstract
The proposed work is to develop and implement field deployable digital twins with intelligent agents within a naval installation Zonal Electrical Distribution System (ZEDS), capable of autonomous learning, decision-making and actuation during #edge case# events and scenarios. The digital twin enables continuous adaptation to events and changing conditions in the field to improve ZEDS resiliency. The digital twin is comprised of a physical twin (the system) and a virtual twin (a model of an asset interacting with the system). This work is part of a collaborative initiative with Naval Post Graduate School and EXWC to develop a building block-based ZEDS approach to naval installation microgrids, where the building block is a nanogrid or nanogrid sub-station. Various nanogrid architectures will be assessed and down-selected, applying a Model-Based Systems Engineering approach. The ZEDS aligns electrical zones of protection with damage control zones and divides power generation, energy storage and loads into self-sustaining geographical zones with redundant sourcing of power/energy between zones during and after damage events. ZEDS also includes autonomous means for isolation of damage between and within zones to mitigate loss(es) of service to undamaged zones and critical loads within damaged zones. Intelligent agents within the ZEDS, such as a Battery Energy Storage System (BESS) and a nanogrid controller, will be proactive in responding to survivability events. For a given asset, a combination of Unsupervised and Supervise Machine Learning will be applied to inform implementation of an Artificial Intelligence plus ML (AI/ML) algorithm that will ultimately be deployed with that asset into the end use ZEDS installation to enable autonomous learning, decision-making and actuation. The digital twin development process proceeds as follows: In the training stage, the virtual twin is a state space model of the nanogrid asset, and the physical twin is a detailed full simulation of the system. In the development stage, the virtual twin is a Controller-Hardware-in-the-Loop (CHiL) model of the asset plus AI/ML (informed by the training stage) that is co-simulated with the physical twin (a detailed Real Time emulationof the system). In the deployment and commissioning stage, the virtual twin plus AI/ML is implemented in a hardware/firmware platform embedded within the deployed asset hardware into the end-use installation. This project will demonstrate all stages of digital twin training, development, and deployment. An approach to sustainment, maintenance and re-commissioning is also suggested as part of a broader, external concurrent project to improve the operational success of microgrids and ZEDS based microgrids.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jan 24, 2024
- Source ID
- N000142412070
Entities
People
- Robert Cuzner
Organizations
- Office of Naval Research
- United States Navy
- University of Wisconsin System