Adaptive on-line tracking of power system harmonics using ADALINE

The active power harmonic filtering is performed by injecting equal-but-opposite of the distortion into the power line. The harmonic on-line tracking is an essential part of the filtering process. In this paper, the linear adaptive neurons (ADALINE), a version of ANN (Artificial Neural Network), is used to perform adaptive on-line tracking of the power system harmonics. The ADALINE can not only accurately estimate the dynamic harmonic amplitudes but also adaptively track the dynamic fundamental frequency in the power system. Moreover, we also proposed the adaptive learning rate for bringing the faster convergence.