The optimum mother wavelet for travelling wave fault locators

Research into travelling waves used for fault location mainly utilize digital transient models of the power system under fault conditions. Electromagnetic transient analysis is the foundation for travelling wave-based fault location estimation, and when they are applied with precise time synchronization, extremely high accuracy is achieved. Most travelling wave-based fault locators detect the time of arrivals of the fault incident surges at both ends of a transmission line, and this requires a level of time precision compatible with the speed of light. Fault signals are non-stationary and non-periodic, and consequently, an important tool to provide multiresolution analysis of the fault signals is the wavelet transform. Wavelets from Daubechies family are normally recommended by researchers that studied faults generated by electromagnetic simulators. However, the real-world travelling wave fault data described in this paper was obtained from the recording of the actual faults on transmission lines; using synchronized fault locators positioned at both ends of various transmission lines. With decomposition and reconstruction of signals at the scaling level of up to six, this paper provides a thorough comparison between the Daubechies, Biorthogonal, Symlet and Coiflet wavelet families, and underlines their accuracy and reliability for use in fault location analysis.