Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

AI-Driven Phase-Shifted Carrier Optimization for Cascaded Bridge Converters, Modular Multilevel Converters, and Reconfigurable Batteries

Amin Hashemi-Zadeh, Nima Tashakor, Sandun Hettiarachchi, Stefan Goetz

TL;DR

This work tackles ripple and voltage distortion in PSC-PWM for modular converters under unbalanced modulation indices. It introduces a compact neural-network surrogate that predicts optimal phase-shift angles from modulation vectors, trained offline with GA-labeled data to enable real-time online control. The method achieves up to a 50% reduction in current ripple and WTHD and is 100k–500k× faster than GA, with a scalable partitioning strategy that reuses trained models for larger systems without retraining. Validation across simulations and hardware experiments on a four-module reconfigurable battery demonstrates robustness to faults and unbalanced conditions, underscoring the approach’s practicality for online modulation optimization in large-scale modular converters.

Abstract

Phase-shifted carrier pulse-width modulation (PSC-PWM) is a widely adopted scheduling algorithm in cascaded bridge converters, modular multilevel converters, and reconfigurable batteries. However, non-uniformed pulse widths for the modules with fixed phase shift angles lead to significant ripple current and output-voltage distortion. Voltage uniformity instead would require optimization of the phase shifts of the individual carriers. However, the computational burden for such optimization is beyond the capabilities of any simple embedded controller. This paper proposes a neural network that emulates the behavior of an instantaneous optimizer with significantly reduced computational burden. The proposed method has the advantages of stable performance in predicting the optimum phase-shift angles under balanced battery modules with non-identical modulation indices without requiring extensive lookup tables, slow numerical optimization, or complex controller tuning. With only one (re)training session for any specified number of modules, the proposed method is readily adaptable to different system sizes. Furthermore, the proposed framework also includes a simple scaling strategy that allows a neural network trained for fewer modules to be reused for larger systems by grouping modules and adjusting their phase shifts. The scaling strategy eliminates the need for retraining. Large-scale assessment, simulations, and experiments demonstrate that, on average, the proposed approach can reduce the current ripple and the weighted total harmonic distortion by up to 50 % in real time and is 100 to 500 thousand times faster than a conventional optimizer (e.g., genetic algorithms), making it the only solution for an online application.

AI-Driven Phase-Shifted Carrier Optimization for Cascaded Bridge Converters, Modular Multilevel Converters, and Reconfigurable Batteries

TL;DR

This work tackles ripple and voltage distortion in PSC-PWM for modular converters under unbalanced modulation indices. It introduces a compact neural-network surrogate that predicts optimal phase-shift angles from modulation vectors, trained offline with GA-labeled data to enable real-time online control. The method achieves up to a 50% reduction in current ripple and WTHD and is 100k–500k× faster than GA, with a scalable partitioning strategy that reuses trained models for larger systems without retraining. Validation across simulations and hardware experiments on a four-module reconfigurable battery demonstrates robustness to faults and unbalanced conditions, underscoring the approach’s practicality for online modulation optimization in large-scale modular converters.

Abstract

Phase-shifted carrier pulse-width modulation (PSC-PWM) is a widely adopted scheduling algorithm in cascaded bridge converters, modular multilevel converters, and reconfigurable batteries. However, non-uniformed pulse widths for the modules with fixed phase shift angles lead to significant ripple current and output-voltage distortion. Voltage uniformity instead would require optimization of the phase shifts of the individual carriers. However, the computational burden for such optimization is beyond the capabilities of any simple embedded controller. This paper proposes a neural network that emulates the behavior of an instantaneous optimizer with significantly reduced computational burden. The proposed method has the advantages of stable performance in predicting the optimum phase-shift angles under balanced battery modules with non-identical modulation indices without requiring extensive lookup tables, slow numerical optimization, or complex controller tuning. With only one (re)training session for any specified number of modules, the proposed method is readily adaptable to different system sizes. Furthermore, the proposed framework also includes a simple scaling strategy that allows a neural network trained for fewer modules to be reused for larger systems by grouping modules and adjusting their phase shifts. The scaling strategy eliminates the need for retraining. Large-scale assessment, simulations, and experiments demonstrate that, on average, the proposed approach can reduce the current ripple and the weighted total harmonic distortion by up to 50 % in real time and is 100 to 500 thousand times faster than a conventional optimizer (e.g., genetic algorithms), making it the only solution for an online application.

Paper Structure

This paper contains 19 sections, 19 equations, 11 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Applicable Modular Topologies: Reconfigurable Batteries as a Case Study
  3. Circuit Description
  4. PSC-PWM and Problem Description
  5. Current Ripple and Harmonic Analysis
  6. Proposed Framework
  7. Neural-Network Modulation and Scheduling Framework
  8. Sample Generation
  9. Cost Function Design and Feature Selection
  10. Comparison Between GA Optimization and Exhaustive Search
  11. Neural Network Training
  12. Expandability to Other Topologies, Module Counts, and Carrier Waveforms
  13. Applicable Circuit Configurations
  14. Neural Network Reuse via Modular Partitioning
  15. Carrier Waveform Mapping: Triangle to Sawtooth
  16. ...and 4 more sections

Figures (11)

  • Figure 1: Generic reconfigurable battery system: (a) topology; (b) module type; (c) module operating modes.
  • Figure 2: PSC-PWM features and performance: (a) shifted carriers; (b) modulation rule; (c) performance under balanced and unbalanced operating conditions for a case study of four modules.
  • Figure 3: Workflow of the proposed framework.
  • Figure 4: The proposed NN integration with control scheme.
  • Figure 5: The proposed method expandability.
  • ...and 6 more figures