Typhoon disasters pose a severe threat to the security of power distribution networks and transportation networks. Pre-positioning mobile emergency power supplies (MEPS) in an optimized manner is a key measure to enhance distribution network resilience

contributing to improved efficiency of post-disaster emergency power dispatch and faster load restoration. To this end

a disaster scenario generation method based on conditional Wasserstein generative adversarial network with gradient penalty (cWGAN-GP) is proposed

and a two-stage stochastic optimization model is constructed to determine the optimal MEPS locations within predetermined areas. The cWGAN-GP is employed to capture the nonlinear relationship between meteorological factors and line failures

thereby generating diverse failure scenarios

from which representative scenarios are extracted using hierarchical clustering. After considering the post-disaster restoration process

a two-stage stochastic mixed-integer programming model integrating power and transportation networks is formulated

and the progressive hedging algorithm (PHA) is applied to solve the optimal MEPS pre-positioning plan for each transit station. A coupled system comprising the IEEE 33-bus distribution network and a 44-node transportation network is used for case studies. The results demonstrate that the proposed optimization method can be adopted to significantly improve MEPS response speed and substantially reduce theload restoration time after disasters.