DEVELOPMENT OF TRAVELLING WAVE PROTECTION ALGORITHMS FOR POWER TRANSMISSION LINES

Abstract

Travelling Wave Protection (TWP) utilises the fault-generated high frequency transients to determine the fault location. It is always desired in power systems as it can provide ultra-high-speed fault clearance, which is recognised as an effective method of increasing power transfer and improving the transient stability of a power system. In this thesis, the detailed analysis of the malfunction of the correlation function-based TWP algorithm is presented, which is currently the most important TWP algorithm. Starting from this, two improved algorithms are proposed to enhance its performance, especially under cases of high resistance faults. Moreover, primary attempts are also presented to develop new TWP algorithms using the latest Wavelet Transform (WT) theory. In the first improved algorithm, an auxiliary correlation function is proposed on the basis of the analysis of the travelling wave propagation. The auxiliary function takes advantage of the travelling wave surges that cause the malfunction of the traditional correlation function. It performs correctly under high resistance faults, which makes it a complementary of the standard correlation function. An adaptive strategy is proposed to select the correct result from the two functions. Hence the algorithm can work reliably under both low resistance and high resistance fault cases. The malfunction of the standard correlation function under high resistance fault is resolved satisfactorily. However, for power systems with weak sources, the performance improvement of the aforementioned algorithm is limited. A peak enhanced correlation algorithm is then proposed to deal with such cases. In this method the pattern recognition technique is employed to collaborate with the correlation function. The peak intensity is enhanced significantly, which improves its noise rejection ability and performance under high resistance faults. In the second half of this thesis, the wavelet transform (WT) theory is introduced to the travelling wave protection. The wavelet transform is an efficient signal analysing tool which has been applied successfully in many fields in the past decade, and its applications in power system have also been reported recently. The correlation function-based travelling wave protection algorithm is analysed from the view point of the wavelet transform theory, which reveals that it can be reformulated as a simple wavelet transform. Moreover, a kind of spline wavelet and the corresponding wavelet transform-based edge detection method are employed to replace the traditional correlation function. The WT-based edge detection algorithm can attain similar performance to the correlation function method, with significant improvement in the computational efficiency. One of the eminent advantages of wavelet transform is that it can extract the feature of the signal through multi-scale wavelet transform. In this thesis, three kinds of multiscale WT are examined, i.e., the Multiresolution Analysis (MRA), the Wavelet Transform-based Singularity Detection (WTSD), and the Wavelet Correlation Function (WCF). Among them, the WCF is especially suitable for the travelling wave protection problem. The TWP algorithm based on WCF exhibits better accuracy, peak sharpness and noise rejection capability than the traditional correlation function method. In conclusion, two improved correlation function-based algorithms and two wavelet transform-based algorithms have been developed in this thesis. Simulation studies have shown their promising prospects for travelling wave protection of the transmission lines.