Study Of The Distribution Network Reconfiguration And DGs’ Maximum Accommodating Capacity

Distribution network reconfiguration is an important means of optimal operation for distribution networks, which is mainly compromised of static optimal reconfiguration and fault recovery reconstruction of distribution networks. Static optimal reconfiguration of distribution networks is an optimizing technology through opening and closing switches to reduce network power loss or balance load. As technology of distribution network reconfiguration is mature, it has become one of the important tool in distribution network planning and distributed power location. In this paper, we study distribution network reconfiguration problems and maximum accommodating capacity of DG to be installed in changeable topology. The thesis consists of three parts:(1) Firstly, in order to improve the convergence precision and computation efficiency of distribution network reconfiguration, this paper presents a topology expression called tree hierarchical encoding method which can reflects relationship between node and node, node and branch. In this method, different types of nodes information are marked on the elementary tree by signs, which can be better applied in the judgment method based on elementary tree topology, thus enhancing the operating efficiency of the topology judgment; Secondly, this paper proposes a topology identification method called separated-trees transforming method, which can transform separated-trees and reconstruct them into the topological solution. As a result, the method can save more time than that based on traditional traversal method and improve computation efficiency. Finally, in order to process infeasible solutions produced in iterative process of intelligent optimizing algorithms, this paper explores an infeasible solution mechanism, which can make corresponding process according to different types of infeasible solution. This mechanism can effectively keep evolution information of infeasible solutions and improve solution’s quality in evolution process.(2) To improve the applicability of particle swarm optimization used in discrete problems, this paper presents an improved algorithm which combines the set-based updating strategy with dynamic-multi-swarm particle swarm algorithm. To further enhance algorithm’s local search ability, a neighborhood learning strategy is presented. By comparing the results with other algorithms’, it shows that the improved algorithm can better balance global search and local search, and can achieve more accurate results in shorter iterations, which prove the superiority of the improved algorithm.(3) In order to explore DGs maximum accommodating capacity in distribution networks with the influence from changeable topology, this paper uses a calculation method based on sensitivity-matrix. In this method, related elements from the Jacobi matrix are extracted to form the sensitivity-matrix. The sensitivity-matrix can reflect the influence of DG’s active power and reactive power on node voltage change. The whole process can be completed through a power flow calculation and matrix operations, which can improve calculation efficiency greatly. Then, an optimization algorithm is applied and the sensitivity-matrix computation method is used as the evaluation function of the optimization algorithm. The maximum of DGs accommodating capacity in changeable topology is calculated and the results show that the changes of topologies greatly influence the maximum accommodating capacity of DGs.