Optimization approach to dynamic restoration of distribution systems

The problem of fault restoration in distribution systems is a multi-objective dynamic combinatorial problem with topology constraints. The problem consists in finding both (1) an admissible post-fault configuration and (2) the sequence of switching operations to reach such configuration. Research on the problem solution has been focusing on the search for admissible post-fault configurations; the sequencing of switching operations is usually disregarded. However, the sequencing of switching operations determines the admissibility of the intermediate configurations and therefore the effectiveness of the overall restoration plan. In this paper, we propose to address the problem of fault restoration in distribution systems in a two phase approach. In the first phase a network optimization evolutionary approach is proposed to find the optimal post-fault configuration. In the second phase a dynamic programming approach is proposed to determine the optimal sequence of switching operations. The two phase approach is illustrated with an example from a real urban distribution network.