Fault Detection in Transmission Line Using Empirical Mode Decomposition and High Speed Decision Tree

Transmission lines as part of electrical power system when experience any fault may cause power interruption in the system. Conventional transmission protection systems face many difficulties. So, classification of fault and identification of its location in transmission lines are very important. So, any fault should be detected very quickly using an appropriate classifier and the faulty phase should immediately be isolated with the help of a proper relaying system. This paper proposes a signal decomposing technique to decompose the original fault signal into different intrinsic mode functions (IMFs). The most notable intrinsic mode function having more resemblances with original fault signal is considered for feature extraction, then a time series-based high speed decision tree is implemented for detection of fault. The considered network is simulated in MATLAB/Simulink platform and simulation result shows appreciable accuracy in varying fault conditions.