Distributionally Robust Resilient Unit Commitment Supported by Battery-Electric Locomotives Under Decision-Dependent Uncertainties

Battery-electric locomotives (BELs) offer a promising solution for long-distance energy exchange in power systems. To address the challenge of random transmission line failures under uncertain routing before hurricanes, this paper proposes a proactive resilience-enhancing scheduling strategy that adeptly integrates unit commitment (UC), energy storage plants, and BELs. The problem is formulated as a two-stage distributionally robust optimization (T-DRO) model with risk aversion. To manage decision-dependent uncertainty (DDU) arising from BEL routing decisions and distributional ambiguity caused by limited outage information during hurricanes, an event-based decision-dependent ambiguity set (DD-AS) with $L_{1}$ distance is developed, which captures random failures in both transmission lines and railway networks to construct a decision-dependent sample space. Using Lagrangian dual method, the optimization problem is reformulated and decoupled into deterministic master problems and subproblems. An enhanced Benders decomposition algorithm with hybrid cuts techniques is designed to solve the proposed T-DRO under event-based DD-AS. The advantage of this algorithm over classical Benders decomposition method is theoretically interpreted. The proposed model and algorithm are tested on the modified IEEE reliability test system (RTS) under a hurricane to verify their effectiveness.