Battery-swap stations reduce the replenishment time of electric vehicles and have considerable regulation potential. An accurate load prediction model is the key to their participation in grid auxiliary services. To address the stochastic nature of the power exchange demand of users, a fuzzy clustering-Markov chain-based load prediction model for swap stations is established. First, the Poisson distribution is used to predict the number of EVs demanding power exchange at each moment and establish the demand constraint for power exchange. Second, the adaptive fuzzy C-mean clustering algorithm is used to adaptively partition the battery clusters in the swap stations based on the charge state to avoid the subjectivity of artificial partitioning. Finally, the Markov chain is used to establish the battery cluster model of the swap stations under multiple states of charging, discharging, and waiting. The demand prediction method and load prediction model are simulated, validated, and compared with the Monte Carlo simulation method. The results showed that Poisson distribution accurately predicted the demand quantity of electric vehicles, and the proposed load prediction model obtained the power of the swap station under charging and discharging states, while reducing the volatility of load prediction.

A two-stage, recovery strategy for distribution networks pre- and post-disasters is proposed to address the issue of transmission-line failures caused by extreme events (such as extremely cold weather and typhoons) which result in a significant loss of power in the distribution network, considering the coordination of battery-charging switching stations (BCSS) for electric vehicles. We analyzed the feasibility of the BCSS participation in the pre-to post-disaster distribution-network restoration and two-stage, restoration architecture and constructed a refined energy-transfer link model for BCSS; a thermal-management system was constructed considering the impact of a low-temperature environment. We proposed a two-stage, recovery strategy for a distribution network that considers the BCSS before and after a disaster. In the pre-disaster stage, the allocation of battery packs was decided. In the post-disaster stage, based on the remaining resource power-supply potential—the objective function was the maximization of the overall economic optimization to ensure a stable load operation. Finally, numerical simulations were conducted using the improved IEEE33 node and 123 node-distribution networks as examples. The results indicated that the proposed strategy is both feasible and superior and could improve the recovery rate of important loads in the distribution network while ensuring lower recovery costs.