AI for Stability Optimization in Low Voltage Direct Current Microgrids

Low voltage direct current (LVDC) is an enabling technology to foster a sustainable resilient energy supply. LVDC microgrids comprising energy generators, storage systems, and loads work as independently controlled units in connection with common alternating current networks. Precise digitized control applying intelligent power converters enables new AI-based approaches for DC microgrid layout and operation. In this work, a new method involving connected machine learning and optimization is established together with a novel measurement system, which enables the measurement and improvement of microgrid stability. The application is successfully validated by experimental assessment on a testbed with a four-terminal DC network operating at a voltage of 380 VDC and the advantages of the AI-based approach are demonstrated.