Multi-Objective Online Optimization Technique of a DC Microgrid

The increasing demand for efficient and sustainable energy systems has intensified research in advanced energy management for DC microgrids. This paper presents an online optimization strategy that continuously updates offline-optimized power references to mitigate discrepancies between predicted and actual system conditions. Unlike conventional standalone online optimization techniques, the proposed method introduces an intrasample online optimization layer that performs local, one-step-ahead updates within each sampling interval. The objective is to minimize both operational costs and toxic gas emissions, subject to operational constraints that ensure the safe and standard performance of the microgrid. The strategy is evaluated against a baseline offline approach that employs the utility grid converter as a slack bus for corrective actions. Validation is performed through a detailed simulation environment under realistic operating scenarios. Results show that the proposed method effectively adjusts the power mix in real time, enhancing the participation of dispatchable sources while achieving superior trade-offs between cost efficiency and environmental impact compared to the offline technique.