Abstract:

A novel anti-aliasing wavelet packet transform method for harmonic detection is proposed. Aiming at the low measurement precision and poor robustness which exists in the former traditional wavelet methods for lack of the aliasing reduction scheme, an optimal interpolation wavelet packet filter is designed according to new optimal criteria. First, the limitation of anti-aliasing on the traditional wavelet filter bank is analyzed. Second, the designed optimal interpolation filters are denoted, and then the solution algorithm is given. This devised wavelet packet filter can seek a reasonable balance between signal preservation and aliasing reduction; it overcomes the inherent bug of traditional wavelet transforms, which rooted from just only concerning total aliasing cancellation but not aliasing-reduction in decomposition. Simulation and several comparative results indicate that the proposed method can effectively eliminate aliasing and precisely extract harmonic information.