Distribution Network Reconfiguration Research

As modern distribution networks have been developing fast, the network is increasingly complex, the makeup of power consumer has changed greatly, and consumer's requirement for improving power quality and supply reliability is more and more stronger, it is difficult to keep a large-scaled distribution network operating safely and economically only by manual management, and even difficult to satisfy the imperative requirement of power utilities for reducing losses and enhancing service quality.Distribution network reconfiguration is an efficient measure to enhance power network's operating efficiency, supply reliability and power quality. It is a general question covered with many aspects, which can be used to different operating task, such as losses reduction, load balancing, reaction optimization, service restoration, and so on.Topology analysis is an indispensable part of network reconfiguration. The reconfiguration algorithm needs to change the network configuration constantly through topology analysis to search optimal network configuration with satisfying objective functions. So the approach of topology analysis is fundamental to the algorithm of distribution reconfiguration, the performance of reconfiguration algorithm is determined to the sound model of network and efficient topology analysis. In this paper, an object-oriented and spatial-oriented model for distribution network is designed. Based on this model, an efficient method of topology analysis is presented which can trace the change of the distribution network automatically and fast. The principle of this algorithm is connectivity analysis based on spatial. Through the analysis computed models and data structures for distribution network reconfiguration algorithm is built up, which can lessen the work force of network modeling and improve the performance of reconfiguration algorithm.Afterwards, the dissertation reviews different algorithm based on differentoptimal technology and different objective functions, and then, the research is emphasized mainly on the reconfiguration algorithm based on minimization of loss and algorithm based on service restoration.The distribution reconfiguration problem is a non-linear, multi-constraints, mixed-integer optimization problem. Though so many optimal techniques, Genetic Algorithm is applied widely in distribution reconfiguration algorithm for its outstanding performance of search optimization. So this dissertation's focus is mainly on the application of Genetic Algorithm to distribution reconfiguration algorithm. This paper presents an improved solution for distribution network reconfiguration based on a refined hybrid genetic algorithm. In the algorithm the "the integer permutation" encoding is adopted with each integer representing one controllable switch. A decoder is designed to decide the final network configuration corresponding every chromosome. A local search operator is combined with the genetic algorithm, which to a certain extent improve the local optimal capability of the algorithm, quicken the algorithm's constringency and improve the algorithm's performance.In normal operating state, network reconfiguration can be undertaken to improve operation efficiency, losses minimization and load balancing. When the system is out of service, the reconfiguration is used to find a restoration solution and implemented to provide as much supply as possible to the customers until the system is returned to the normal state.The network reconfiguration based on distribution service restoration is significant for power utilities to enhance supply continuity, consumer satisfactory, and reduce the losses of consumer and utilities. In the process of real-word service restoration, because of nonstandard network plan, restriction of line and transformation capabilities, it is unavoidable to cross voltage levels to look for more spare capacity in the system.In this paper, taking account into the status quo and features of practical distribution network, a HV-MV coordination restoration algorithm is propos...