Localization and Imagination of Line-like Soft Faults in Transmission Lines

Soft fault detection for transmission lines is of great significance to secure electricity distribution network, since a soft fault will gradually develop into a hard fault or electrical fires. The existing detection methods are often applied to localize point-like faults, while most of soft faults are line-like faults with a certain length. Considering that soft faults will change the signal propagation speed, this paper establishes a transmission line model with respect to speed distribution. Then, a wavenumber domain reflectometry (WNDR) is developed to obtain the speed distribution using Volterra inverse scattering series (VISS). By dealing with zero wavenumber compensation and Gibbs ringing artifact suppression issues of VISS in the wavenumber domain, WNDR achieves not only localization but also imagination of fault length and dielectric permittivity for line-like faults, without the need of time synchronization and the knowledge of line topology. The simulation results show that WNDR locates the range of line-like faults more precisely than the existing fault detection methods, and achieves dielectric permittivity imagination with an accuracy larger than 97%.