A comparative study between wavelet primarily based ANN and ANFIS algorithm technique to locate fault in a transmission line

In this paper; a comparative study approach is proposed between Feed Forward Neural Network (ANN) and accommodative Adaptive Neuro-Fuzzy Inference System (ANFIS) rule model to find the fault of 3-ph transmission line using MATLAB Toolbox. Post fault current is the key requirement of the technique and its features are extracted using DWT `DB4' as mother wavelet. These features are then used to train the ANN and ANFIS models. The training to testing ratio is maintained at 7:3 for both the algorithm models. Two test cases on the basis of number of training samples are considered for analysis for each type of four faults viz.; BC; ABC; ACG; AG. ANFIS algorithm proved to be the better algorithm approach as compared; to efficiently estimate the fault location with an extreme standard of accuracy and at minimum requirement of dataset samples.