Discretized Energy Control

Abstract

This project proposes to build upon recent developments in distributed predictive control utilizing networked Universal Controllersfor PEBB-based systems. Recent developments on the Universal Controller have shown that high speed communication between FPGA-basedcontrollers can be performed with latencies compatible with the lowest level feedback controls. This type of networked power electronics system is capable of energy flow control in a much faster and more dynamic manner than possible when system control is performed at higher control layers, as it has been to date.The research proposed here will further the development of distributed network-based coordination of PEBBs to enable more dynamic energy flow control. Methods of fast predictive coordination will be explored in order to direct energy along desired pathways by moving it into and out of specified PEBBs and groupings of PEBBs. The ultimate goal of the proposed research is commanding movement of energy in intervals correlated to the control and communication timing so that all applications exist exclusively as data flows on the PEBB communication network. Thus, re-programming of converter functions would no longer be needed for various applications. Energy patterns would be manipulated directly in real-time via the network such that application-based energy flow patterns could be dictated on a control cycle timing basis. Dynamic adjustment of energy flows via network traffic within physical constraints will yield an amorphous power system.In order to accomplish the overall project goal severalsteps must be performed. First, currently used model-predictive control methods for power electronics will be extended to increase robustness against uncertainties that result in model mismatch. Methods to account for PEBB operational limits dynamically will alsobe investigated. This is needed for creation of precise energy packets. Next methods for representing power electronic converter functions in terms of energy flows that are then mapped as network data flows will be developed. To accomplish this an appropriate representation of the energy packets using graph theory will be investigated. Finally, in order for the energy packet control frameworkto be applied to existing analysis techniques and interface standards as well as synthesizing various filtering functions it shouldbe related to the concept of impedance or to immittance more generally. Extension of frequency domain quantities used to calculate immittances for non-periodic quantities will be investigated.

Document Details

Document Type
DoD Grant Award
Publication Date
Feb 02, 2021
Source ID
N000142112093

Entities

People

  • Herbert L. Ginn

Organizations

  • Office of Naval Research
  • United States Navy
  • University of South Carolina

Tags

Readers

  • Computer Networking
  • Robotics and Automation.
  • Theoretical Analysis.

Technology Areas

  • Directed Energy
  • Microelectronics