Transmission Line Fault Location Using MFCC and LS-SVR

The location of Transmission Line (TL) Faults is a major problem in Electrical Power Systems (EPSs), since precisely identifying the point occurrence fault TL it possible to perform faster restoration operation desired normal conditions. In this work we used Least-Squares Support Vector Regression (LS-SVR) locate faults with inputs provided by MFCC (Mel-Frequency Cepstral Coefficients) obtained from voltage signals during fault. A modelled line based on parameters real was used, total 4008 situations being simulated Line. It important out that are not applications involving EPS’s, and, according bibliographic research conducted team so far, no application feature extraction tool has been detected for problem. 3006 were train model cross-validation k-fold method, and 1002 testing. proposed methodology presented good performance tests carried out, mean relative error 0.000419 ±0.000640% when models trained tested noiseless signals. For present SNR ranging 100 dB 25 dB, ranged 0.00334 ±0.00459%, first case, 0.030580±0.043160% last.