Security-aware Coordinated Dispatch of Microgrid: an Adaptive Online Optimization Approach

Kaidi Huang,Lin Cheng,Ning Qi,Qiang Ji,David Wenzhong Gao

DOI: https://doi.org/10.1109/pesgm51994.2024.10688696

2024-01-01

Abstract:Renewable energy sources (RES) integrated into microgrids (MG) exhibit high volatility and limited predictability, posing challenges in the reliable and economic dispatch of the MG. Contrary to the standard approaches that require predictions of RES and load, this paper proposes a novel prediction-independent two-stage coordinated dispatch method, where the day-ahead dispatch employs scenario-based stochastic optimization solved by a simultaneous backward reduction method, while the intra-day corrective dispatch uses the online convex optimization (OCO) framework. To achieve the tractable and efficient solution of OCO with inter-temporal constraints, an adaptive virtual-queue-based algorithm is newly introduced that exhibits sublinear dynamic regret and sublinear hard cumulative constraint violation. Case studies illustrate that the proposed method outperforms state-of-the-art methods in terms of economic optimality, computational efficiency, and security, regardless of operational scenarios.