Extreme learning machine based fault classification in a series compensated transmission line

This paper presents an accurate hybrid method for fault classification in a series compensated transmission line. The proposed scheme uses discrete wavelet transform in combination with extreme learning machine for fault classification. Instantaneous value of current signal is measured from the relaying end of the transmission line for one cycle duration from the inception of fault. Discrete wavelet transform is used to decompose the signal and extract certain features from it. The feature set is then normalized and best features are selected from the total feature set by forward feature selection method. Selected features are then fed as an input to the extreme learning machine for fault classification. To evaluate the feasibility of the proposed technique, it is tested on a 400 kV, 300 km series compensated transmission line for all the ten type of fault using MATLAB/ SIMULINK. A wide range of simulation condition is taken to generate the train and test pattern. Simulation result indicates that the proposed approach is robust, fast in learning and classifies the fault very accurately.