Research on Multi-source Cooperative Operation Strategy of Urban Distribution Network Based on Data-driven Distributionally Robust

The interaction between electric vehicle charging stations(EVCS) and PV, energy storage, and other equipment progressively increases as the penetration rate of electric vehicles and distributed new energy increases. Simultaneously, the safe operation of urban distribution network will be significantly impacted by the volatility and randomness introduced by new energy access. Consequently, the issue of examining the multisource cooperative operation strategy of urban distribution network in light of uncertainty has emerged as a pressing requirement for urgent research and resolution. To address the uncertainty of PV prediction, the data-driven algorithm of the Wasserstein generative adversarial network is employed to generate PV output scenarios. PV typical scenarios are obtained through K-mediods clustering cuts, and the probabilistic confidence intervals of the distribution of the typical scenarios are constrained by the 1 -paradigm and $\infty$-paradigm numbers. Subsequently, the typical scenario data of PV is employed to develop a data-driven distributionally robust urban distribution grid multi-source cooperative operation model that takes into account electric vehicle charging stations, energy storage, and PV. The first stage of this model ascertains the power procured from the primary grid, whereas the second stage identifies the PV power relinquished, the storage charging and discharging capacity, and the equivalent load of the EVCS. Finally, the suggested model’s validity is confirmed using data from an urban distribution network.