Abstract
In the process of distribution network fault recovery, in order to better address the issues caused by the uncertainty of new energy power output, this paper proposes a multi period power supply recovery method for distribution networks based on adaptive data driven approach. Firstly, this method uses the historical data of new energy power output to construct an ellipsoidal uncertainty set, and forms a data driven convex hull polyhedral set by connecting the vertices of the high-dimensional ellipsoid. Then, aiming at the problem of large conservatism during the reduction process of the convex hull polyhedral set, based on the range of the box set, cutting planes are made starting from the vertices, and a data driven hyperplane polyhedral set model is constructed. Furthermore, considering the constraints of cyber-physical integration, an adaptive data driven power supply recovery model for distribution networks is established, and the column and constraint generation (C&CG) algorithm is adopted to solve the robust scheduling model. Finally, simulations on the improved IEEE-33 bus system and actual example systems verify that the adaptive data driven power supply recovery model for distribution networks can reduce conservatism and improve the robustness of the optimization results. This power supply recovery model can reduce conservatism and enhance the robustness of the optimization results. In the actual distribution network fault recovery scenarios, it can make more efficient use of new energy and ensure the stability of power supply, which strongly demonstrates the effectiveness and application value of the proposed method.