Non-Stationary Harmonics Analysis Based on Wavelet Packet Transform and Neural Network
The method for analysis of stationary harmonics in power system is FFT, but it is unsuitable for non-stationary harmonics. Because of the feature that non-stationary harmonics’ frequency spectrum has a certain bandwidth and with some noise interference usually. A new method for detection, based on wavelet packet transform and neural network was presented in this paper. This method improved the traditional wavelet analysis method. The non-stationary harmonics were decomposed in different frequency bands by wavelet packet transform at first, and then complete the analysis of the non-stationary harmonic in different frequency bands. Through software simulation, the analysis results show that, the method has better accuracy, and provided an effective means for analyzing non-stationary harmonics.