Multiphase interleaved buck converters offer high efficiency and low ripple but face significant instability caused by constant power loads (CPLs) and various disturbances in DC microgrids. This paper proposes a novel control strategy based on generalized predictive control (GPC) integrated with a current balancing mechanism. By utilizing the dynamics of total inductor current and output voltage, the controller significantly reduces computational complexity compared to conventional model predictive control (MPC). To optimize transient performance, the GPC prediction horizon is adaptively tuned in real-time using a Takagi–Sugeno fuzzy inference system, effectively minimizing overshoot and settling time. Furthermore, a fixed-time disturbance observer is incorporated to estimate uncertainties and disturbances in DC microgrids without requiring prior knowledge of initial conditions or additional sensors. Comparative studies against conventional PI, standard MPC and non-adaptive GPC in MATLAB/Simulink demonstrate that the proposed method provides superior dynamic response and enhanced robustness against CPL-induced instability. Finally, processor-in-the-loop tests on an STM32 Nucleo-G431RB microcontroller validate the practical feasibility and real-time effectiveness of the strategy for industrial applications.
