Application of Adaptive Neural Networks in Harmonic Current Detection

A novel harmonic current detection method based on the adaptive neural networks is proposed in this paper to improve the performance of active power filter(APF)'s harmonic curent detection.According to the basic principle of adaptive noise cancelling technology,the harmonic current is obtained after the fundamental current is filtered from the load current.This method can accurately and fast detect harmonic current and well overcome the shortages of those detection methods based on FFT,the instantaneous reactive power theory and the wavelet transform.Simulation results based on MATLAB/Simulink testify the real-timing and accuracy of the method.So the method proposed in this paper can be used as the harmonic current detection of APF.