A Graph-theoretic Framework for Electric Power Distribution System Service Restoration

Distribution system restoration (DSR) aims at restoring the service after a fault to the maximum number of customers while ensuring the satisfaction of critical system operating constraints. DSR is essentially a system reconfiguration approach that entails determining a combination of open and closed switches to obtain a new radial network topology that optimizes certain performance criteria related to service restoration. Given the combinatorial nature of the associated search problem, the existing literature either use heuristic methods or present non-linear mixed integer models for DSR, both computationally intensive for a large distribution system. This paper presents a graph-theoretic distribution system restoration strategy that maximizes the restored load by formulating the restoration problem as a mixed-integer linear program (MILP) while incorporating critical system connectivity and operating constraints. The novelty lies in variable selection and transformation that help incorporate integer decision variable with the continuous power flow variables. Simulation results presented using IEEE-37 node distribution system and a 1069-node multi-feeder test case demonstrate the effectiveness of the proposed approach.