Wavelet-Based Fault Location on Power Transmission Lines using Travelling Waves

Abstract

The invention of travelling wave fault locators has increased the confidence in accurately locating faults on transmission lines, as compared to other techniques. The travelling wave fault location (TWFL) methods are mostly studied using simulated models of actual and idealised power systems. Regrettably, access to real-world travelling wave (TW) fault data is rare in the field of power system protection, and throwing tests on actual transmission networks are almost unknown. Researchers focus on utilising simulators, or running experiments in laboratories, to investigate the behaviour of a power system when a short-circuit fault occurs on a transmission line. Their outcomes are used to demonstrate the accuracy of TW fault locators, but their reliability and accuracy on various occasions have been a significant challenge, especially when the cause of a fault is lightning. The travelling waves instigated by a short-circuit fault are often complicated, and some TWFLs do not always deliver reasonable results. The well-known signal processing methods used in conventional TWFLs are correlation, cross-correlation, pattern recognition, STFT-based fault locator and wavelet-based methods. Real-world travelling wave fault data was provided for the practical examinations in this PhD study by an industrial collaborator. So, data analysis became a major part of this research; and this facilitated data cleansing and categorisation. Thus, the practical contribution of the study is to introduce a scientific and practical well-developed fault location method based on real-world data, capable of being implemented in a commercial product and utilised globally. This thesis furthermore contributes to the science and industry by examining the mother wavelet functions that have been commonly used in the field and proposes the use of a Biorthogonal wavelet transform as the most reliable wavelet for TWFL. Finally, disturbances occur on transmission lines, and they were seen in the recorded data. They cause maloperation of fault locators, especially the time-domain methods. This thesis introduces a disturbance detection algorithm, as the third contribution, by using statistical and state-of-the-art digital signal processing techniques. The final contribution of this PhD project is to publish articles and present posters in conferences, journals and seminars globally. To increase the likelihood of commercial success with the proposed TWFL, the Engineering and Physical Sciences Research Council (EPSRC) and the industrial collaborator have funded an Impact Accelerator Award (IAA) for a six-month Proof-of-Concept project. This facilitates the collaboration between The University of Manchester and the company to implement the proposed fault location method in a commercial product.

Date of Award	28 Jan 2021
Original language	English
Awarding Institution	The University of Manchester
Supervisor	Haiyu Li (Main Supervisor) & Patrick Gaydecki (Co Supervisor)