Wavelet-Based Directional Protection Supported by Wavelet-Based Fault Classifier
Directional protection, fault classifier, wavelet transform
Fast and accurate protections are required to ensure the safety of the power system. In
this research, a new directional protection based on wavelet transfom is described by using
only the first wavelet decomposition level. The scaling coefficients are used in replacement of
the Fourier transform in order to recreate the conventional directional protection. The torque
equations were mathematically redefined in the wavelet domain by using sampled voltages and
currents. Regarding the protection activation, the wavelet coefficients are used for fast detection of fault-induced transients (wavelet activators) in order to reduce the relay operating time,
whereas the scaling coefficients are used for backup activation in accordance with the overcurrent protection (wavelet overcurrent activators). In addition, a real-time fault classification
method using the RT-BSWT applied to modal components of the Clarke transform was developed to support the directional protection. The method uses a simple threshold-based logic
flow instead of a sophisticated artificial intelligence-based algorithm, and requires three-phase
current measurements from only one terminal. Furthermore, by using the wavelet coefficients
energy of RT-BSWT, the method is fast and accurate due to the extraction of both low- and
high-frequency components of faults. Evaluations have proved the feasibility to rebuild the
conventional directional protection by using the RT-SWT, which has provided better performance and faster operating time than the conventional protection based on the discrete Fourier
transform (DFT). As new functionalities to the directional protection, the wavelet-based negative sequence unit can be used to identify the fault directionality to three-phase faults, even with
severe voltage sags without memory strategies, which is not possible with the conventional protection. The fault classifier results have shown it is accurate and fast to identify all of the ten
fault types successfully supporting the directional protection needs