An iterative inversion method for transmission line fault location

This dissertation discusses various transmission line forward modeling techniques in both time and frequency domains. Although time domain methods offer simplicity in most cases, the computational inefficiency and lack of fidelity make these methods less attractive. Therefore, the more efficient frequency domain technique is emphasized - a modified transmission matrix (also known as ABCD) method. One of the most difficult problems in electrical wire fault location nowadays is detecting and locating frayed wiring, where the wire is only partially damaged. This type of fault can be very small and extremely difficult to detect. Most inversion schemes used to locate faults require forward models that accurately represent detailed reflections. Resolving these very small faults requires an especially accurate forward model where not only the fault but also all the other very small changes caused by normal aspects of the wiring system are included. A very high resolution Finite Difference Time Domain (FDTD) method can be used to simulate this type of fault and details of the surrounding wiring system with enough fidelity to distinguish the small fault. However, this is very costly in terms of computational resources. This dissertation demonstrates a quick way of building the fray profile that significantly reduces the simulation time. Finally, the ultimate goal of the highly realistic forward modeling is the inversion, in which a set of measured data is given and the inversion algorithm interprets the location and the nature of fault on the wire. Multiple iterations are typically required, and thus, high efficiency is necessary. A new method introduced in this dissertation is capable of identifying multiple unknown parameters in just a few steps.