Intercircuit and cross-country fault detection and classification using artificial neural network

The conductor geometry in double circuit lines makes them prone to multi-circuit faults like earthed and unearthed intercircuit faults and cross country faults. The probability of inter-circuit faults is increased when the multiple lines are mounted on the same tower. Mutual coupling is also present during un-earthed intercircuit faults. The phase-to-phase inter-circuit fault (without earth connection) provokes the presence of zero-sequence current, which is detected by ground distance relays. The consequences of intercircuit faults are often not considered in conventional relay design philosophies. In this paper both earthed and un-earthed intercircuit faults are investigated. Artificial neural network based technique has been employed for detection and faulty phase identification (classification) of intercircuit and cross-country faults. The study is carried out on a MATLAB® platform and results of ANN based fault detector/classifier are presented and discussed. The simulated test results shows that this technique detects and identifies the faulted phase correctly and quickly over wide range of power system operating conditions; which is a striking benefit of ANN based technique.