Robust model predictive control of a four-switch buck-boost converter for fast dynamic response with guaranteed ZVS

Abstract

Although four-switch buck–boost (FSBB) converter provides high efficiency, compact structure, and zero-voltage switching (ZVS) capability, its dynamic performance and robust-ness under parameter variations remain insufficiently explored. This paper proposes a unified continuous-control-set model predictive control (CCS–MPC) framework that incorporates ZVS-oriented modeling, stability-based control parameter design methodology, and adaptive disturbance compensation. A hybrid compensation scheme is proposed, combining an extended state observer (ESO) with a least-mean-square (LMS) adaptive filter. This scheme enhances disturbance rejection and mitigates the bandwidth–stability trade-off in conventional ESOs. Experimental results validate that the proposed control achieves steady-state convergence within several switching cycles, maintains full ZVS with 98.5% efficiency, and preserves low computational complexity comparable to PI control. These results confirm that the proposed CCS–MPC offers a practical solution for high-efficiency, fast-response DC–DC converters operating under model inaccuracies.