Study Of Faulted Cable Line Selection And Fault Location

With the need of the modern city development, at present, more and more the middle-low voltage power supply networks transport the electric power from trolly wire to electric power electric cable. But, when the electric cable used under the ground break down, it is very difficult to find the fault because of the cable complicated, so this has blocked the development of reliability of power supply and automation level of distribution network. On the other hand, cable faults result in directly the economy loses and affect people's life. Therefore, the study about fault line selection and location has practical urgently significance. This paper first deeply analyzes stable and transient features of single phase-to-ground fault in power distribution network in china. Both theory analysis and experiment results show that there are many high frequency component features in transient fault currents. Wavelet analysis is a new stage of signal analysis. On one hand, it overcomes Fourier Transform that is not equal to analyzing unsteady signal and transient signal; On the other hand, it overcomes Fourier Transform that the scale of window of time domain and frequency domain is fixedness. It can analyses Multi-resolution signal and break signal. Therefore, the faulted line is detected in terms of the features of magnitude and polarity of wavelet coefficient of transient zero-sequence current in certain scale, thus the proposed method overcomes the limitation of steady and other transient approaches. Simultaneously, on the base of fault line selection, wavelet packet is used to analyze the transient signal, and to find that is connection with transient fault signal and fault distance. Because artificial neural network has the nonlinear approximation, the paper realizes fault distance measurement algorithm with wavelet packet and artificial neural networks. Great deal of simulation data show, the proposed method of cable fault location by using wavelet neuron network can online select the faulted line and measure the fault distance quickly and exactly in middle-low voltage power supply networks.