Research On Reliability Evaluation And Network Reconfiguration Of Distribution System

The reliability of distribution system greatly influences reliability of the customers, because of the distribution system directly connecting to the customers. Distribution system reconfiguration might find an optimal operation pattern of reducing the network loss, improving power quality and supply reliability by changing the status of the tie or sectionalizing switches, under the constrains of radialized operation, the balance between power supply and consumption, capacity and voltage limit, and etc.. The characteristic of the distribution system is that it owns large quantities of components, devices, and the very complex network. Therefore, how to evaluate the distribution system reliability and find an optimal solution for distribution system reconfiguration rapidly and accurately are two important issues for long time. Aimed at the above problems, the contents and achievement of this thesis are as follows:1 In this thesis, a new reliability evaluation method based on network simplification for distribution system with the consideration of the switch fault is presented. The method divides the one-line diagram of the system into zones with the auto (or non-auto) switch to be the boundary firstly; then two simplified equivalent networks are respectively formed by two types of zone-elements (Type A or Type M), which are defined according to the type of the switches in the zones; a valid proposition is suggested and proved for the simplification of the equivalent networks. The switch is looked as one of components of the relevant zone-element to avoid the consideration of the effect of switch failure independently. Since the zone-load and the zone-element are treated as the basic evaluation subject and analysis unit respectively, and a classification evaluation strategy adopted for the reliability indexes calculation, the method proposed has the advantages of reducing the number of the units to be ransacked, avoiding repeated evaluation of the same type of load points; thus large improvement of the computational efficiency compared with the other methods is achieved. Test result of the samples demonstrates the effectiveness and validness of the method.2 Adaptive Immune Algorithm (AIA) based on Immune Algorithm (IA) is proposed to apply to the distribution system reconfiguration for the first time. By automatically adjusting the parameters: selecting rateα, cloning radius r and mutation radius R with the help of the measure of distance between the antibodies, and simultaneously searching in both larger and smaller regions, AIA achieves the balance between the fast convergence and diversity of population, which escapes from the local optimum. Based on the heuristic rule, a decimalization coding method is designed, which reduces the complexity of the problem to a certain extent. The test results show that AIA owns obviously advantage comparing with other methods, such as Genetic Algorithm (GA) and Immune Algorithm (IA).3 A new method called AINOCOP(Adaptive Immune algorithm for Numerical Optimization of COnstrained Problem)which combines AIA with the feasible solution searching and co-evolutionary strategy of GENOCOPⅢ(Genetic algorithm for Numerical Optimization of COnstrained Problems) to repair the infeasible solution for the distribution network reconfiguration is further introduced. The AINOCOP makes use of the advantage of Bi-level optimization of AIA to accelerate the calculation, and at the same time, the co-evolutionary method of GENOCOPⅢis used to correct the infeasible solutions to be the feasible solution; therefore, the search space of AIA is expanded, the diversity of the population is maintained, and the efficiency and precision of the calculation for the AIA is further improved. The test results show the effectiveness of the proposed method AINOCOP.