Fault analysis in transmission lines using neural network and wavelets

Transmission lines can be viewed as link between generating stations and the consumers. They are the major part of the power system hence they play a vital role in its operation. Being exposed to atmosphere they are quite vulnerable to any type of fault. Faults are a transient phenomenon and result in high frequency content in the fault voltage and current signals at related nodes. They add a high frequency spectrum to the original signal at the instant of fault. Analysis of this high frequency spectrum can be done using discrete wavelet transform. It is observed that pattern of the spectrum containing a band of frequency changes with initiation and type of any disturbance. It also depends on the location of fault in the transmission line. While taking the energy factor, the negative going signal is also accounted instead of having a cancellation effect and thus energy factors of various voltage signals serve as effective inputs to the neural network to classify the fault.