Service Restoration in the Medium Voltage Grid Minimizing the SAIDI Contribution After Primary Substation Failures

One of the most widely used reliability measures for the electric grid is the System Average Interruption Duration Index (SAIDI), i.e. how many minutes the average customer is without power per year. Primary substation failures may significantly increase SAIDI, since they affect a large number of customers. This may be avoided by improving the service restoration process, which requires solving complex non-convex combinatorial problems. Hence, this paper proposes an algorithm for smart, SAIDI-optimizing service restoration in meshed MV grids after primary substation failures. We first formulate service restoration as a Mixed-Integer Linear Programming (MILP) problem by using a linearization of the power flow equations that is proper for meshed grids and by applying exact Big M relaxations for linearizing other involved non-convexities. The decisions include among others the travel routes of the field crew and account for the grid security constraints. Then, we propose a low complexity heuristic iterative method that optimally solves the MILP problem. The results for a test case in Switzerland suggest that the SAIDI contribution can be reduced by more than 10% compared to the currently deployed restoration strategy at the utility. Also, they show the effectiveness of the SAIDI objective versus common objectives in the literature.