Dual-mode model predictive control for non-minimum phase boost converters

Abstract

This paper aims to develop an efficient finite-set model predictive control (FS-MPC) strategy for DC-DC boost converters to improve voltage regulation while reducing computational complexity. The proposed approach introduces a split cost function that decouples voltage and current regulation, providing a simpler alternative to conventional long-horizon FS-MPC schemes used to address the converter’s non-minimum-phase (NMP) behavior. A current estimation technique is incorporated to eliminate the need for additional sensors, lowering hardware cost and improving robustness. Unlike existing FS-MPC methods that rely on horizon extension or extra measurements, the proposed strategy leverages the split cost structure to achieve comparable NMP compensation with significantly lower computational effort. The controller is implemented in real time using a hardware-in-the-loop (HIL) setup on a ZedBoard platform, with accurate data acquisition provided by an external ADC. Experimental results demonstrate that the proposed approach enhances voltage-tracking performance, eliminates overshoot and undershoot, reduces settling time by over 40%, and decreases computational effort by more than 80% compared to traditional FS-MPC methods.