A Heuristic Approach Applied to Time Reversal MUSIC Method for Soft Fault Location in Noisy Transmission Line Networks

Abstract

Time-Reversal multiple signal classification (TR-MUSIC) has emerged as a promising technique to locate multiple soft faults in complex wire networks, thanks to its location accuracy and sub-millimeter resolution. The surprising ability to operate using continuous wave excitations even at low frequencies makes its readily adaptable to live wire testing. However, real-life networks are always surrounded by different sources of noise, an issue which is rarely addressed in the domain of fault detection and location. Intrinsically, TR-MUSIC capable of resolving multiple scatterers in open media has shown resolution degrading problems in the presence of perturbations. Accordingly, we will study in this paper the effect noise can bring on the fault location accuracy in different complex wire networks. This will be followed by proposing a heuristic clustering approach which will enable returning good estimates of the fault location at signal to noise ratio (SNR) well below 5 dB.