Research On The Reactive Power Optimization Of Distribution System With Distributed Generation

As a flexible power supply mode, distributed generation plays an important role in slaving distribution network investment and reducing power loss and so on through combining with distribution network. At the same time, distributed generation influences in the distribution system in many aspects, such as distribution network planning, power flow and active power loss and so on. This paper studies on software implementation for the reactive power optimization of distribution system with distributed generation.First, back/forward sweep algorithm is researched deeply and a practical implement method of back/forward sweep distribution power flow is proposed. The topology structure of distribution network is defined by the tree structure which consists of special circuit branch structure and node structure. There is no need to form admittance matrix in this implement method. The validity and accuracy of the implement method above is verified through a practical example.Second, the influence brought by different connection position and connection capacity of distributed generation is analysed. The assumption that total capacity of distributed generation should be no more than the total capacity of power load is proposed. Furthermore, to the software implemented by C language which finds a reasonable connection position and connection of distributed generation is discussed.The assumption which total capacity of distributed generation should be no more than the total capacity of power load is verified.Last, the reactive power optimization of distribution system after the connection of distributed generation is studied. Based on genetic algorithm, main procedures of software implementation are analyzed and the accuracy of the software implementation is verified though an example. Furthermore, a software of reactive power optimization used C programming language is implemented and the validity and accuracy of it is verified through the statistic analysis of many experiments.