enhancing frequency stability in islanded microgrids via model predictive control

This paper proposes a model predictive control-based strategy for secondary load frequency control to enhance the dynamic performance of such systems. the proposed controller generates optimal control signals for dispatchable units to minimize frequency deviations induced by load and generation variability. a comprehensive microgrid model is developed, incorporating photovoltaic arrays, wind turbines, fuel cells, battery and flywheel energy storage systems, diesel generators, and electrolyzers. the dynamic behavior of each component is formulated using small-signal transfer functions, and the mpc is designed based on a constrained quadratic optimization problem that predicts and mitigates frequency deviations. simulation results in matlab/simulink demonstrate the superiority of the proposed mpc approach compared to conventional and intelligent controllers, including ziegler–nichols tuned pi, fuzzy-pi, cpso-pid, and cpso-fopid. the proposed controller achieved a maximum frequency deviation of 0.0052 pu, a settling time of 5.1 seconds, and an itae of 0.00024—outperforming all benchmarks in both steady-state and transient scenarios. robustness under system parameter variations and load disturbances was also validated through five distinct case studies. the controller exhibits improved reliability, reduced stress on primary controllers, and better resilience to uncertainties. future work will focus on implementing adaptive mpc algorithms, integrating machine learning-based disturbance predictors, and validating the control scheme using real-time hardware-in-the-loop platforms for enhanced applicability in hybrid ac/dc microgrids.