Radial basis neural network and discrete wavelet transform based RMS estimation for fault assessment in three phase transmission line

Quick identification, accurate classification and proper estimation of location of faults in transmission line are utmost required for smooth and reliable operation of electrical system. This work deals with the line-to-ground (LG) fault assessment in transmission line. At one end three phase current signals have been analyzed by discrete Wavelet transform (DWT) based root-mean-square (rms) analysis for fast detection of faults. Then computed rms values have been used in Radial Basis Neural Network (RBNN) for classification of faulty phase and estimation of fault locations. Using the proposed technique, very quickly fault has been detected and with high accuracy almost 100% fault has been classified and distance of fault properly estimated. In different cases algorithm have been tested successfully and the results are very much optimistic.