Multi-objective Model And Genetic Algorithm For Service Restoration Of Distribution Network

The distribution system links directly to customers whose reliability and power quality are affected by the operation of distribution system. The restoration of the distribution network is an optimization problem that the out-of-service area need be restored after the fault location and isolation. How to restore the out-of-service area effectively is significant to decrease outage loss, shorten outage time and improve supply reliability.Based on the currently research on domestic and international literatures of distribution service restoration, the main works are as follows:â‘ A multi-objective model of distribution service restoration is presented: in the objective function, minimization of load rejection ranks, minimization of load rejection capacity, load balance index, minimization of switch operation and minimization of the losses are taken into account. The advantage of this model lies on taking multi-objective function into consideration completely.â‘¡Conventional genetic algorithms for distribution network reconfiguration are unsatisfactory because they are easy to result in infeasible solutions. To remedy such a situation, necessary and sufficient conditions for radial run of simplified network which directs two-stage coding and decoding method are proposed Firstly, the simplified network's links considered as the opened branch-groups are searched via the minimum spanning tree algorithm-Prim algorithm. Secondly, the opened branches in the branch-groups that have been selected in the first step are determined. Throughout the entire algorithm process, no infeasible solution comes out.â‘¢Based on bee evolutionary genetic algorithm single-objective reconfiguration of distribution network is studied which makes a foundation for multi-objective reconfiguration of distribution network. Using the proposed algorithm, it can avoid falling into premature convergence and can effectively improve search performance.â‘£Study the problem of multi-objective service restoration of distribution network Firstly a two-stage service restoration method is presented to solve the multi-objective model of distribution service restoration. In the first stage, the objective function includes load balance index, minimization of switch operation and minimization of the losses. In this case, multi-objective reconfiguration of distribution network is conducted. In the second stage, the objective function includes minimization of load rejection ranks, minimization of load rejection capacity and minimization of switch operation. In this case, load rejection is conducted because feeder line current limits or voltage limits are violated. For solving the service restoration problem a fast and elitist non-dominated sorting genetic algorithm-II (NSGA-II) based on two-stage coding and decoding method is proposed. In this method,both reliability and economy of the restored network are considered. In addition,each run of the algorithm can produce several Pareto-optimal solutions. Thus, several solution programs can be selected by decision-maker.This method is implemented and applied into IEEE 16-bus,69-bus and a practical network. The calculation results indicate that the present algorithms are correct and effective.