A New Method for Measurement of Harmonic Distortion Using Adaptive Wavelet Neural Network

Abstract: The proliferation of power electronic based nonlinear loads and time varying devices causes harmonic pollution in industrial power system in recent years. The harmonic distortion can cause overheating and increased losses in the equipments used in distribution system and also interference with the communication systems. This paper presents a new soft computing technique based on an adaptive wavelet neural network (AWNN) for harmonic distortion measurement. Wavelet Neural Network (WNN) is a new technique recently proposed for harmonic distortion monitoring. In this work, Mexican hat wavelet has been selected for activation function in the hidden layer of the network. The validation of proposed AWNN is examined with feed forward back propagation network (FFBPN). The proposed method has been verified that the improved estimation accuracy and low computational time, when compared to the FFBPN