Pre-disaster distributionally robust scheduling model for active distribution networks considering topology reconfiguration

Extreme natural disasters represent particularly significant challenges to the safe and economical operation of electric distribution networks (DNs) utilizing a high proportion of intermittent renewable energy. However, although topological reconstruction is regarded as an ideal approach to maximize the capability of DNs in rapidly restoring normal load supply in the face of emergencies, the randomness of fault occurrence and the uncertainty of renewable energy output have been relatively overlooked in previous research. The present work addresses this issue by proposing a pre-disaster distributionally robust scheduling model for active DNs considering topology reconfiguration with a high proportion of photovoltaic (PV) and energy storage system (ESS) resources while seeking to optimize the energy sources, electric power grids, electric loads, and ESSs (i.e., source–grid–load–storage coordination) collaboratively under PV output uncertainty. The model is a two-layer model. First, the upper-layer model aims to obtain the optimal DN topology that minimizes the total dispatch cost in the pre-disaster prevention stage under multiple fault scenarios using stochastic programming. Second, the lower-layer model applies a distributionally robust optimization (DRO) approach to optimize the source–load–storage scheduling strategy based on the DN topology given by the upper layer model, combined with the PV output scenarios. The DRO problem is solved using the column and constraint generation algorithm (C&CG). Finally, the effectiveness of the proposed two-layer scheduling scheme for improving the load survivability of a DN under extreme weather conditions and PV output uncertainties is verified based on the computational results obtained for an IEEE 33-bus system as an example.