Current Constrained Single-Loop Model Predictive Voltage Control for DC-DC Synchronous Buck Converter using Online Optimization

For the voltage control problem of DC-DC synchronous buck converters, the dual-loop proportional-integral (PI) control scheme is typically used. Despite the simple structure, optimal PI gains for the dual-loop PI voltage control could not be obtained easily. Moreover, the outer-loop PI voltage control deals with the constraint of inductor current by adding the anti-windup term which also required some tuning efforts. In this paper, to cope with the current constraint and reduce the number of control loops, a single-loop current-constrained model predictive voltage control (MPVC) for the buck converter is proposed. A cost function for the voltage control problem is constructed using a one-step-ahead state error. The current constraint is represented in the optimization problem by using the control input and the discrete-time system model. The quadratic programming (QP) solver using the Hildreth's algorithm is implemented to generate the optimal duty cycle for the buck converter. Comparative simulation results between the proposed MPVC and the typical and anti-windup dual-loop PI controllers are presented.