Optimal energy restoration in radial distribution systems using a genetic approach and graph chain representation

Abstract Several approaches have been developed to restore energy in distribution systems after the interruption of services for part of the system. In order to achieve the energy restoration plan, these approaches usually perform the restoration for circuits with specific features. Additionally, they are not able to deal with large outage zones. This paper develops a method using Genetic algorithms for energy restoration in radial distribution systems. This algorithm overcomes the combinatorial characteristic and deals with non-linear and non-continuous objective functions, inherent to this kind of problem. A new representation for distribution systems is also proposed aiming to avoid the generation of unfeasible configurations (with outage zones and/or closed loops). The generation of only feasible configurations saves RAM memory and processing time. The efficiency of the proposed approach is shown using a fairly complex distribution system.