Stability Analysis for Model Predictive Voltage Control of Buck Converter

Model Predictive Control is a modern method of nonlinear control which gives superior dynamic performance with the cost of the increased computational burden. In this paper, a continuous control set model predictive control (CCS-MPC) strategy for a buck converter is proposed, which also limits inductor current within specified limits. The control strategy achieves a constant switching frequency with the PWM modulator, and its stability is analyzed. The proposed algorithm is based on the sampled data model of the buck converter, which is simplified using a polynomial approximation to reduce the computational complexity. The paper also shows how this approximation is valid for most practically designed buck converters and analyzes the computational requirement. The strategy is improved by limiting the duty ratio within the stability region. The computation is implemented within a short time with the help of hardware accelerators like floating-point unit (FPU) and trigonometric math unit (TMU) in the Delfino microcontroller from Texas instruments. The proposed control strategy is verified in simulation and compared with experimental results, and it shows good performance for both reference tracking and disturbance rejection. It is about five times faster than conventional PI with a lead controller.