Real-Time Energy Management for Hybrid AC/DC Microgrids with Adjustable Bus Voltage

Studies on dynamic voltage adjustment at the power electronics level within DC microgrids are prevalent, yet there is a scarcity of research on corresponding energy management strategies. This article introduces a novel approach that integrates DC bus voltage as an adjustable variable within the dispatch model, which incorporates a DC/DC and DC/AC converters power loss model. This model is real-time solved using the Deep Deterministic Policy Gradient (DDPG) algorithm, enhanced with a corrector. The strategy enables dynamic adjustment of the low-voltage DC bus in AC-DC microgrids, significantly reducing losses in DC/DC and DC/AC converters, thereby lowering operational costs. Simulations conducted in Python validate the effectiveness of this method, demonstrating its potential to improve energy management in microgrids.