Publications (3)0 Total impact
Conference Paper: Wavelet-based partial discharge denoising using hidden Markov model[Show abstract] [Hide abstract]
ABSTRACT: Wavelet-domain hidden Markov models (HMMs) have recently been introduced and applied to signal and image processing. The advantage of the method is that the HMMs measure the dependency between the wavelet coefficients and have no free parameters in denoising. In this paper, the HMMs method is applied in reducing partial discharge (PD) white noise. The effectiveness of the method is demonstrated by using numerical simulations and real-world data of neutral point current of generator. Compared with the shrinkage method, the result shows that the HMMs method is better in enhancing signal-to-noise ratio and reserves more PD impulses.
Conference Paper: Partial discharge denoising via lifting scheme[Show abstract] [Hide abstract]
ABSTRACT: Denoising is a key technique in partial discharge detection and analysis. In this paper, a new wavelet construction method, named the lifting scheme, is applied to depress the white noise. A biorthogonal wavelets arc is designed via the lifting scheme which has adaptive characteristics (signal independence). Hard thresholding is used in signal denoising. Numerical experiments show that this algorithm has a better de-noising effect than other methods due to its better balance between noise removal and PD pulse attenuation.
- [Show abstract] [Hide abstract]
ABSTRACT: According to the characteristics of fault transient signal of a power system, some principles of fault signal compression are proposed. That is keeping some waveband lossless compression and other waveband lossy compression. An integrated signal compression algorithm based on wavelet packet transform, Huffman coding and vector quantization is proposed. Programming software is also given. The simulation result shows this algorithm can precisely meet the requirements of signal compression in a power system and have great application potential.