Research On Fault Location Method In Distribution Network

Single-phase grounding fault has the greatest probability of all the faults occur in small current grounding system. Fault location timely is of great significant to maintaining the stable operation of the power system. However, fault feature is difficult to extract because of the complex topology of distribution network and noise interference, especially the arc suppression coil makes the fault signal too weak to detect. For these reasons, the problem of fault location in distribution network has been unable to obtain a better solution. After the fault signal de-nosing, the fault features which are applied to fault location on the next step will stand out more clearly.In this paper, a complete ensemble empirical mode decomposition with adaptive noise(CEEMDAN) method is presented to the de-noising signal using the decomposed difference of white noise and useful signal. And the difference is applied to identify the demarcation point bet ween the noise IMF function and useful IMF function by a threshold. Although high-frequency IMF include noise signal substantially, removing it directly may result in lossing the high frequency useful signal. In this paper, the wavelet soft threshold de-noising is applied to those Mode Function(IMF) to acquire the useful information in high-frequency. In the end, all the useful Mode Function(IMF) signals will be reconstructed. Compared with some De-nosing method such as the wavelet analysis and ensemble e mpirical mode decomposition(EEMD), it can be concluded that the new method here presented do not need to select any basis function, and has better performance in de-noising and adaptability.There is a correlation exists between typical frequencies of fau lt signals and specific paths in the network covered by the travelling waves. Fault features can be extracted clearer after signals de-noised. However, the voltage transient will be attenuated due to the reflection or refraction at the wave impedance disco ntinued. This paper illustrates a procedure based on the new transient signal analytical tool—spectral kurtosis method for the analysis of voltage transient due to phase-to-ground fault. All-phase FFT—a method used to identified the fault line section through comparing the phase angles of the measured currents at two ends of a line section. And a threshold value is presented to determine whether there is a fault. The end of the line section is selected as a observation point to obtained the line-mode component of fault voltage waves. With the typical frequencies of fault signals analyzed by spectral kurtosis, the fault location can be calculated accurately.