An Offset-Free Composite Model Predictive Control Strategy for DC/DC Buck Converter Feeding Constant Power Loads

Qianwen Xu, Yunda Yan, Chuanlin Zhang, Tomislav Dragicevic, Frede Blaabjerg

Research output: Contribution to journalJournal articleResearchpeer-review

3 Citations (Scopus)

Abstract

The high penetration of power electronic converters into dc microgrids may cause the constant power load stability issues, which could lead to large voltage oscillations or even system collapse. On the other hand, dynamic performance should be satisfied in the control of power electronic converter systems with small overshoot, less oscillations, and smooth transient performance. This article proposes an offset-free model predictive controller for a dc/dc buck converter feeding constant power loads with guaranteed dynamic performance and stability. First, a receding horizon optimization problem is formulated for optimal voltage tracking. To deal with the unknown load variation and system uncertainties, a higher order sliding mode observer is designed and integrated into the optimization problem. Then an explicit closed-loop solution is obtained by solving the receding horizon optimization problem offline. A rigorous stability analysis is performed to ensure the system large signal stability. The proposed controller achieves optimized transient dynamics and accurate tracking with simple implementation. The effectiveness of the proposed controller is validated by simulation and experimental results.
Original languageEnglish
Article number8839850
JournalI E E E Transactions on Power Electronics
Volume35
Issue number5
Pages (from-to)5331-5342
Number of pages12
ISSN0885-8993
DOIs
Publication statusPublished - May 2020

Keywords

  • Constant power load (CPL)
  • large signal stability
  • Model Predictive Control (MPC)
  • nonlinear disturbance observer
  • offset-free tracking

Fingerprint Dive into the research topics of 'An Offset-Free Composite Model Predictive Control Strategy for DC/DC Buck Converter Feeding Constant Power Loads'. Together they form a unique fingerprint.

Cite this