A Lightweight Polynomial Regression Controller for Sustainable Grid-Connected DC Microgrids with Enhanced Voltage Regulation

The transition toward sustainable energy systems requires reliable, efficient, and computationally practical control strategies for renewable energy-based microgrids. Grid-connected DC microgrids provide an effective platform for integrating distributed renewable energy resources, while their sustainable operation requires robust regulation under load variations, nonlinear loads, and input disturbances. This study proposes a lightweight Polynomial Regression Controller (PRC) for voltage regulation in grid-connected DC microgrids. The proposed data-driven controller uses a second-order polynomial model to estimate the converter duty cycle from input voltage, voltage error, and load current. The model is trained offline using independently generated operating trajectories and evaluated under previously unseen operating conditions. The results demonstrate accurate DC bus voltage regulation and robust operation under linear, constant power, motor load, and grid-connected conditions. The PRC maintains the DC bus voltage close to its 50 V reference, with steady-state errors of 0.002–0.008% and a settling time of 0.001 s under fast transient responses. The proposed approach combines nonlinear mapping capability with a compact computational structure, supporting practical implementation on resource-constrained platforms. Overall, the proposed PRC contributes to reliable renewable energy integration, resilient microgrid operation, and the development of sustainable, efficient, and scalable smart energy systems.