DWT-Based Investigation of phase currents for Detecting High Impedance Faults Due to Leaning Trees in Unearthed MV Networks

In this paper, features of earth faults due to leaning trees are extracted from the network phase currents using the discrete wavelet transform (DWT). Due to the associated arc reignitions, the initial transients in the electrical network give behavioral traits. The detection security is also enhanced because the DWT is responded to a periodicity of the initial transients. The absolute sum of the DWT detail d3 coefficient corresponding to the frequency band 12.5-6.25 kHz is computed over one power cycle for each phase current of each feeder where the sampling frequency is 100 kHz. It is found that the faulty phase has the highest absolute sum when it is compared with the other healthy phases. Similarly, the absolute sum of the faulty feeder is the highest when the comparison is carried out with respect to other feeders. Therefore, two logic functions are suggested to determine the faulty phase and the faulty feeder. The fault due to a leaning tree occurring at different locations in an unearthed 20 kV network is simulated by ATP/EMTP and the arc model is implemented using the universal arc representation.