A novel Particle Swarm method for distribution system optimal reconfiguration

Searching for the optimal radial configuration of large distribution systems is a problem leading to combinatorial explosion of the number of configurations to analyze. Several deterministic and heuristic methods have been applied in order to endeavor near-optimal solutions. However, no method can ensure that the optimal solution is found in a finite time, so that developing and testing new heuristics is an open and challenging task. This paper illustrates the development of an original Particle Swarm Optimization (PSO) method applied to the minimum losses reconfiguration of large distribution systems. The classical scheme of the PSO has been revisited in order to develop an efficient heuristic, able to take into account the system radial configurations and the various operational constraints of real distribution systems. The resulting PSO method has been successfully applied to large real urban distribution systems. A hybrid formulation including some steps performed with a deterministic iterative improvement method has shown the best performances. The results obtained by using the proposed method on a large real urban MV distribution system are shown and compared to the ones obtained from other methods such as deterministic iterative improvement and simulated annealing.