This paper presents a hybrid model predictive controller to ensure dc microgrid stability and enhance the performance of dc-dc boost converters interfaced with constant power loads in a hybrid system. Hybrid systems are dynamic systems with both continuous current mode (CCM) and discontinuous current mode (DCM) states. In this regard, an automatic model, considering different modes of operation induced by semiconductor switches in dc-dc boost converters and highly non-linear nature of constant power loads is employed to design the proposed control approach. The nonlinear constant power loads connected directly to a dc-dc boost converter is utilized to define an optimal tracking control problem by minimizing a finite-prediction horizon cost function. The proposed controller, which is implemented in both continuous and discontinuous current modes, accounts for the regulation of output voltage within the predefined range. The effectiveness of the proposed hybrid model predictive controller is verified using a comparative evaluation with discrete-time averaged model predictive control (DTA-MPC) and the conventional PI control under experimental conditions. The results authenticate an improved dynamic performance, which can be applied to practical dc microgrids with constant power loads.
