A robust approach for active distribution network restoration based on scenario techniques considering load and DG uncertainties

Fluctuating outputs of distributed generations, time-varying load demands and estimation errors of loads bring substantial uncertainty risks to the active distribution network restoration, which becomes a challenge to the traditional deterministic algorithms. In this paper, a robust restoration approach is proposed to solve this issue, considering both DG outputs and load demands uncertainties. Firstly, a large number of simulation scenarios are generated according to the profiles of historical data. Then we employ the backward scenario reduction technique to cluster these scenarios for computational efficiency. Based on the set of reduced scenarios, the robust restoration control model is built to obtain robust expectedly optimal strategies, which is in the formulation of a mixed integer linear programming. Numerical tests implemented on a modified PG&E 69-bus system demonstrate the robustness and optimality of this proposed approach.