Ambiguity cancellation for wire fault location based on cable life profile

Abstract

Although reflectometry is an efficient method to diagnose simple topologies (such as transmission line, Y shape network), it remains limited in the case of complex branched networks due to multipath fading of the test signal during its propagation. Generally, the knowledge of the environment in which the cable operates gives an additional idea about the fault location. The current paper proposes to introduce the cable life profile (such as environmental stress, type, age, noise, etc.) to detect and cancel diagnosis ambiguities and provide a precise location of the fault. Bayesian Network (BN) seems to be a suitable solution to offer a coherent representation of knowledge domain (reflectometry method, cable characteristics and network heterogeneity) under uncertainties (fault(s) location, systems reliability and measurement precision). In this work, a two-stages BN model for diagnosis using reflectometry in branched networks is proposed and simulation results are discussed.