Technical note Model-based identification of wire network topology Qinghai Shi ⇑ , Olfa Kanoun Chair for Measurement and Sensor Technology, Chemnitz University of Technology, Chemnitz, Germany article info Article history: Received 12 November 2013 Received in revised form 28 April 2014 Accepted 7 May 2014 Available online 24 May 2014 Keywords: Scattering parameters Reflection coefficient Impedance Spectroscopy Global optimization technique Wire fault location Network topology abstract A new technique is proposed to locate wire faults and identify wire network topology using Impedance Spectroscopy (IS). The propagation along the cables is analytically modelled with flexible multi section cascading features utilizing frequency dependent scattering parameters. Therefore, it does not have the common numerical method problems. The transmission line model has the same spectrum as the measured reflection coefficient (q) of wire under test (WUT) so that same practical effects such as skin and proximity effects, signal loss, dispersion and frequency dependent signal propagation can be exactly incorporated. For determination of model parameters an inverse problem should be resolved and differential evolution (DE) approach is proposed. The novel method allows locating hard (short and open circuit) and soft (frays and junctions) faults and also for char- acterization of defects in the branches of network. Results are presented to validate and illustrate the performance of this proposed method. Ó 2014 Elsevier Ltd. All rights reserved. 1. Introduction The increasing use of the wiring in vehicles, communi- cation and power systems has caused growing require- ment for characterization of wire network and location of wire faults. The most widely used technique is reflectome- try. Thereby, a high-frequency signal is send down the cable. The reflected signal including information about changes of cable impedance is used to locate wiring faults. Over the last decade, many methods, such as Time Domain Reflectometry (TDR) [1,2], Frequency Domain Reflectome- try (FDR) [3,4], Ultra Wide Band (UWB) based TDR [5], and Spectrum Time Domain Reflectometry (STDR) [6] were developed. They use different incident signal and signal processing methods. However, these techniques fail to detect soft faults such as frays or chafes and identification of the network topology in practice. Some improved TDR methods [7] use the baseline method, in which the output signal of the faulty wire is compared with the output of the healthy wire, in order to enable to detect and locate the soft faults. The TDR method has the advantage to identify the type of wire faults. Other methods focus more on improving location accuracy. But for all these methods, the wave propagation velocity is approximately considered as a constant parameter which is only dependent on the permittivity of the cable. It limits the accuracy of wire fault location because the time of flight is transformed to the location by means of the wave prop- agation velocity, which is dependent on the wire type and the available frequencies of the excitation signal. Furthermore, soft faults, such as frays and chafes, and multiple faults are difficult to detect by the baseline method. Especially at higher complexity of the wire topol- ogy there are practical difficulties because of noise, multi- ple reflections, unknown load impedances, mechanical variations and changes of electrical parameters due to different wires types. The aim of this study is the development of an auto- mated method for the wire fault location. With this new method the type of wire faults can be identified and the network topology can be characterized. In this study the propagation along the cables is analytically modelled with http://dx.doi.org/10.1016/j.measurement.2014.05.008 0263-2241/Ó 2014 Elsevier Ltd. All rights reserved. ⇑ Corresponding author. Tel.: +49 371 531 37788; fax: +49 371 531 837788. E-mail address: qinghai.shi@ieee.org (Q. Shi). Measurement 55 (2014) 206–211 Contents lists available at ScienceDirect Measurement journal homepage: www.elsevier.com/locate/measurement