Automated wire fault location using impedance spectroscopy and Differential Evolution

Abstract

A new technique is proposed to detect and locate wire fault using impedance spectroscopy (IS) and a model-based global optimization technique. The propagation along the cables is analytical modeled with flexible multi section cascading features by frequency dependent ABCD parameters. Therefore, it doesn't have the common numerical method problems for instance using finite element method (FEM) or finite difference in time domain (FDTD), in which the simulation time is linear or proportional to the wiring length and complexity of the transmission line systems. The transmission line model has been setting as the same spectrum as the measured input impedance of the transmission line system using impedance spectroscopy so that same practical effects such as signal loss, dispersion and frequency dependent signal propagation can be exactly incorporated. Therefore, there is good matching between simulation and measurement data and consequently the optimization technique has a fast convergence and best accuracy. Differential Evolution (DE) is used to solve the inverse problem. The novel method allows locating hard (short and open circuit) and soft (frays, junctions) faults and measuring the impedance of wire fault and consequently the types of wire fault can be distinguished. The developed method is also for characterization of defects in the branches of network. Results are presented to validate and illustrate the performance of this proposed method.