High impedance faults detection in power distribution system by combination of artificial neural network and wavelet transform

In this paper, Discrete Wavelet Transform (DWT) and Artificial Neural Network (ANN) are used to detect High Impedance Faults (HIFs) in Distribution Power System (DPS). In this method, the difference between HIF, Low Impedance Fault and Normal events is established. The DWT techniques are used as pre-processor to de-noise and extract features from three phase current signals at each DPS´s feeder; such features can provide enough information to distinguish by ANN between different fault types. Faulty and normal operations have been simulated by Matlab 6.5. Details of the design procedure and the performance results of proposed method are presented in this paper. The result shows that the proposed method gives good results in detecting HIFs.