A Distributionally Robust Post-Disaster Recovery Method for Distribution Networks Considering Line Repair and Spatiotemporal Dynamic Scheduling of Mobile Energy Storage

Abstract

Extreme disasters often cause large-scale power outages in distribution networks due to damaged lines, significantly impacting system reliability. Current research faces several challenges: traditional methods fail to fully consider the uncertainty of line repair time, existing robust optimization methods encounter difficulties in solving problems involving mobile energy storage (ME), and simple symmetric intervals cannot accurately describe the uncertainty of repair time. To address these challenges, this paper proposes a two-stage distributionally robust post-disaster recovery model that optimizes the connection location of ME in the first stage and adjusts the output of resources such as ME in the second stage to minimize load loss. A Weibull distribution is introduced to fit the repair time of damaged lines, while confidence intervals replace simple symmetric fluctuation intervals to handle the uncertainty of line repair time, improving prediction credibility. The column and constraint generation algorithm is applied to decompose and solve the model. Case studies demonstrate the proposed method's efficacy in maintaining power supply during recovery by simultaneously addressing repair time uncertainty and PV generation variability. At least 26% of the load can be in service even under worst-case scenarios.