Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Experimental Realization of Koopman-Model Predictive Control for an AC-DC Converter

Shun Hirose, Shiu Mochiyama, Yoshihiko Susuki

TL;DR

The paper tackles high-performance control of a nonlinear, time-varying AC-DC converter by developing a data-driven lifted modeling approach using the Koopman operator and Generalized State-Space Averaging (GSSA). It introduces Dynamic Observables to form a lifted state $z[k]$ with $z[k+1]=A z[k]+B u[k]$ and solves a KoMAP-based MPC (K-MPC) to regulate the DC voltage mean and the AC current phasor while enforcing hard current and PF constraints. The method is experimentally validated on a single-phase full-bridge boost rectifier, showing superior steady-state accuracy and transient response compared with IDA-PBC and PI, including adherence to safety/current limits during load transients. This work demonstrates a practical, data-driven control framework for power electronics that can enhance performance in applications like More Electric Aircraft (MEA) and autonomous microgrids. The combination of GSSA-based lifting and MPC offers a robust, constraint-aware alternative to traditional control strategies in nonlinear, time-varying converters.

Abstract

This paper experimentally demonstrates the Koopman-Model Predictive Control (K-MPC) for a real AC-DC converter. The converter is typically modeled with a nonlinear time-variant plant. We introduce a new dynamical approach to lifting measurable dynamics from the plant and constructing a linear time-invariant model that is consistent with control objectives of the converter. We show that the lifting approach, combined with the K-MPC controller, performs well across the full experimental system and outperforms existing control strategies in terms of both steady-state and transient responses.

Experimental Realization of Koopman-Model Predictive Control for an AC-DC Converter

TL;DR

The paper tackles high-performance control of a nonlinear, time-varying AC-DC converter by developing a data-driven lifted modeling approach using the Koopman operator and Generalized State-Space Averaging (GSSA). It introduces Dynamic Observables to form a lifted state with and solves a KoMAP-based MPC (K-MPC) to regulate the DC voltage mean and the AC current phasor while enforcing hard current and PF constraints. The method is experimentally validated on a single-phase full-bridge boost rectifier, showing superior steady-state accuracy and transient response compared with IDA-PBC and PI, including adherence to safety/current limits during load transients. This work demonstrates a practical, data-driven control framework for power electronics that can enhance performance in applications like More Electric Aircraft (MEA) and autonomous microgrids. The combination of GSSA-based lifting and MPC offers a robust, constraint-aware alternative to traditional control strategies in nonlinear, time-varying converters.

Abstract

This paper experimentally demonstrates the Koopman-Model Predictive Control (K-MPC) for a real AC-DC converter. The converter is typically modeled with a nonlinear time-variant plant. We introduce a new dynamical approach to lifting measurable dynamics from the plant and constructing a linear time-invariant model that is consistent with control objectives of the converter. We show that the lifting approach, combined with the K-MPC controller, performs well across the full experimental system and outperforms existing control strategies in terms of both steady-state and transient responses.
Paper Structure (13 sections, 9 equations, 7 figures)

This paper contains 13 sections, 9 equations, 7 figures.

Table of Contents

  1. Introduction
  2. Prototype AC-DC Converter
  3. Controller Design
  4. Overview
  5. Synthesis of the Lifted Dynamics
  6. Introduction to the Input Variable
  7. Model Validation
  8. MPC Formulation
  9. Experimental Results
  10. Conventional Strategies
  11. Results and Discussion
  12. Concluding Remarks
  13. Physics-Based Model

Figures (7)

  • Figure 1: Basic circuit of an AC-DC converter. The Constant Power Load (CPL) is represented by the symbol of current source.
  • Figure 2: Photographs of the experimental system.
  • Figure 3: Comparison of measured data and predictions by the Koopman model.
  • Figure 4: Block diagram of the cascade PI controller.
  • Figure 5: Experimental results of K-MPC (left), IDA-PBC (center), and PI (right). Graphs in row 1 show the DC voltage (blue line) and its reference (red line), graphs in row 2 the AC current (blue line) and the AC supply voltage (red line), and graphs in row 3 the duty ratio. The vertical dotted lines represent $34\,\rm ms$ and $54\,\rm ms$, respectively.
  • ...and 2 more figures