Model Predictive Control of buck DC-DC converter with nonlinear inductor

Abstract

A model predictive control (MPC) approach that deals with the nonlinear behavior of the inductor in the buck DC-DC converter is presented. Based on a discrete approximation of a coil dynamics, a piecewise affine model that better represents the B-H characteristics is derived. For the fast execution of the MPC algorithm, a concept of explicit solutions to MPC is adopted. Two different implementation schemes adapted to very fast sampling rates are outlined, one based on a binary search tree, the other on a polynomial approximation. Provided experimental results show that the proposed MPC scheme better limits the current and importantly, avoids the saturation of the magnetic core.