Electric Ship Charging Load Scenario Generation Method Based on D-W-CGAN

In recent years, port electrification has been gradually and widely promoted, and the charging load of electric ships has surged. The randomness and volatility of the charging behavior of electric ships have brought significant challenges to power grid operation and port load management. Scenario generation simulates realistic charging processes by modeling charging load uncertainties for future port power system planning. A diffusion generation adversarial network is proposed for electric ship charging scenario generation, capable of generating charging station-level electric ship charging time series data with different temporal characteristics. The algorithm draws on the stability architecture of the diffusion model, uses a non-Markov process for forward noise addition, and implicitly models the reverse denoising process using conditional generative adversarial networks. In addition, The superior generative power of the diffuse generative adversarial network proposed in this paper is validated on a real dataset containing many charging session records. The simulation results show that the model proposed in this paper can learn the load characteristics of electric vessels and effectively and accurately generate the set of electric vessel charging load scenarios.