Centralized Self-Healing Scheme for Electrical Distribution Systems

In this paper, a two-stage procedure is proposed in order to solve the centralized self-healing scheme for electrical distribution systems. The considered self-healing actions are the reconfiguration of the distribution grid and, if needed, node and zone load-shedding. Thus, the proposed procedure determines the status of the switching devices in order to effectively isolate a faulty zone and minimize the number of de-energized nodes and zones, while ensuring that the operative and electrical constraints of the system are not violated. The proposed method is comprised of two stages. The first stage solves a mixed integer linear programming (MILP) problem in order to obtain the binary decision variables for the self-healing scheme (i.e., the switching device status and energized zones). In the second stage, a nonlinear programming (NLP) problem is solved in order to adjust the steady-state operating point of the topology found in the first stage (i.e., correction of the continuous electrical variables and load-shedding optimization). Commercial optimization solvers are used in the first stage to solve the MILP problem and in the second stage to solve the NLP problem. A 44-node test system and a real Brazilian distribution system with 964-nodes were used to test and verify the proposed methodology.