Incorporating Data-Driven Uncertainty Set Into Hybrid Microgrids With High Renewable Energy Penetration and Bi-Directional Converters

The energy management of AC/DC hybrid microgrids (HMGs) with high renewable energy resource (RES) penetration and numerous bi-directional converter (BDC) interfaces poses severe challenges due to the output uncertainty of massive RES units and the non-convex characteristics of converter modeling. The present work addresses this issue by constructing a tight data-driven uncertainty set, and implementing non-convex complementary constraints for BDCs. Firstly, an intersected pair convex hull uncertainty set with adjustable confidence levels is developed to closely approximate the original pair convex hull uncertainty set in a computationally efficient way. Secondly, sufficient conditions for relaxing non-convex BDC operational constraints are evaluated, and a checking loop procedure is designed to prevent violations of these constraints. Thirdly, globally optimal solutions of the proposed robust energy management framework for HMGs with RESs and BDCs are obtained via a computationally tractable method based on a parallel iterative algorithm using multiple starting points and implementing the proposed checking loop procedure. The effectiveness, superiority, and computational scalability of the proposed approach are verified on both IEEE 66-bus and large-scale IEEE 246-bus systems.