Vijay S. Kale, Sudhir R. Bhide, and Prashant P. Bedekar
Fault location, mutual coupling, neural network, transmission line protection, wavelet transform
Most of the algorithms developed for the protection of transmission
line use the estimate of the fault location to make the relaying
decision. However, this estimate needs only be accurate enough to
determine whether the fault is in the appropriate zone of protection.
More accurate estimate of the fault location is required to carry
out the repairs, which can save time and resources for the electric
utility. This paper presents an accurate fault location method which
is based on discrete wavelet transform (DWT) and artiﬁcial neural
network (NN). Fault signals were preprocessed with DWT, followed
by the use of appropriate NN, based on the type of the fault, which
gives distance to a fault. The proposed fault location technique
requires signals from one end of the transmission line. Therefore,
it is more reliable as compared to methods which require signals
from both the ends of the line. MATLAB/Simulink software was
used for simulation studies and testing the proposed algorithm.
Large number of faults scenarios were generated by varying fault
resistance, fault location and fault inception time, on a given power
system model. The eﬀect of inter circuit mutual coupling was
considered by choosing a parallel line transmission network. The
results presented on the performance of the proposed technique
clearly demonstrate that NNs are very eﬀective in locating a fault
with a very small error.