Research On Real-time Monitoring Technology For Line Fault In Distribution Network

The main task of distribution network is to provide safe and reliable power supply, however, because of the interferences such as weather, environment, and human factors. Power interruption is frequently caused by short-circuit fault, causing extensive power outages in severe cases. After the failure occurred, as fault current is weak and the arc are unstable, or other reasons, the real-time fault monitoring has not been well resolved long-term. To this day, a number of substations still use manual methods such as pull-Lu fault lines and visual search to find the fault point. The author cayrry out in-depth study in the key area of monitoring technologies, including fault diagnosis, fault detection, fault location, etc. and have a number of innovative research results.First of all, according to the shortcomings of traditional fault diagnosis method, method of distribution network based on RBF neural network is proposed. The method is based on the basic theory of RBF neural network, and combined with a simple model of distribution network, a distribution network fault diagnosis method based on RBF neural network model is proposed. The detailed implementation of the diagnostic process is explained and thus the effectiveness of the method is tested by numerical examples. This method has certain fault-tolerant capabilities, but also is conducive to apply the real-time diagnosis in engineering application.Secondly, on the basis of the existing study, the information fusion model is used to distribution network fault detection. The fault signal is analyzed by wavelet processing, then the evidence of various independent body (wavelet signal representation) is merged using the improved DS evidence theory, making the fusion of each fuzzy uncertain results to better reflect the intrinsic fault signal. Thereby, the alignment algorithm is used to make the final diagnosis to determine the fault phase. Finally, according to the proposed diagnostic theory, the model framework is designed for practical application.Finally, according to the difficulties of finding the fault point in distribution network, a new type of distribution network fault location system is proposed. Firstly, the theory of the distribution network fault location system is described, and the various modules of system implementation are presented in detailn. As a core module, a fault locatgion method based on simulation matching is proposed. Basing on symmetrical component method and linear superposition principle, the additional positive sequence network is established. Then the fault location factor is proposed and matching during-fault recorded factors with the during-fault simulated factors is used to determine the fault location. The search process to find the best waveform match is actually an optimization problem. The simulations carried out in paper justify the reliability and advancement of the method.