Application Study Of Fault Location For Distribution Network Containing Distributed Generation

With the intensive global energy situation and the environment problem being increasingly serious, distributed generation with the feature of low energy consumption, clean environment and flexible power supply have been attached importance and used widely in recent years. A large number of distributed generations access to distribution network have changed the structure and operation mode of distribution network, which brought much influence to the current fault location methods or even made the location methods judge wrong. Solution of the problem of fault location for distribution network with distributed power has great significance for improving the reliability of distribution network and promoting the development of distributed generation.In this paper, the stable and transient features of fault in distribution network with distributed generation are theoretical researched and analyzed. The result of simulation analysis show two conclusions in distribution systems with distributed generation. First, if the branch of positive, negative and zero sequence of fault current in the positive direction that associate with multi-branch node can get the largest integrated amplitude, which would be the fault branch. The second, the transient process of fault also contains a lot of high-frequency characteristics of information that would indicate the direction of fault. Since wavelet analysis has good characteristics of both localized of time-frequency, which can decompose the signal into different frequency bands, so that wavelet analysis were particularly suited to analysis of non-stable signals and transient signals. The zero sequence current of each branch that associates with multi-branch node was simulated by using wavelet analysis. The results of simulation show that the branch in the positive direction that associate with multi-branch node would be the fault only when the transient zero sequence current of modulus maxima of wavelet transform get the largest value and which polarity contrary to the zero sequence of system voltage. According to the fault of the steady-state characteristics and transient characteristics, two criterions were fixed, which could certain the fault place between two multi-branch nodes firstly, then, the accurate fault place can be located between two switches by using matrix algorithm.Genetic algorithm is a global search algorithm, which used in the application of the traditional distribution network to some extent solved the problem of fault tolerance for fault location in distribution network.In order to solve the problem of fault located inaccurately since the signal of fault excessive current uploading by FTU is distortion or loss, the genetic algorithm that has better fault tolerance was attempted to achieve fault location in this paper. The key of fault location based on genetic algorithm in distribution network with distributed generation is determining the direction of each switch and building a new switching function and fitness function. In this paper, the suitable switching function and fitness function was created firstly; then, some example was simulation by using the genetic algorithm in Matlab7.0, the results of simulation show that genetic algorithm can locate fault accurately in distribution network with distributed generation. In the end of paper, a comparison of genetic algorithm and matrix algorithm was used for fault location in the case of signal distortion or loss. The simulation analysis show that genetic algorithm has higher fault tolerance based the ability of global optimization than matrix algorithm.The research of fault location in distribution network with distributed generation still in the initial stage, in this paper, the most commonly used method of fault location based on the matrix algorithm and artificial intelligence were used for, the simulation results of several examples test and verify the effectiveness of the algorithm, which has important theoretical and practical value.The research of this paper is funded by special fund of Key Laboratory of Guangdong Regular Higher Education Institutions under award No.ZDSYS200701and Key Projects of Education Department of Guangdong Province under award No.070043.