In order to accurately detect power grid harmonics in noisy environments,a new method of harmonic signal de-tection is proposed by combining improved wavelet threshold noise reduction and improved complete ensemble empirical mode decomposition with adaptive noise(ICEEMDAN).The method first adjusts the threshold characteristics through the adjustment factorβ,and uses improved wavelet thresholding to pre-process the noise-containing harmonic signals for noise reduction.The processed signal is then decomposed by ICEEMDAN to obtain several intrinsic mode function(IMF)com-ponents,and the spurious components are removed by the correlation coefficient criterion.Finally the amplitude-frequency information is demodulated by the Hilbert transform(HT).Simulation experiments show that the combination of ICEEMDAN and the improved wavelet thresholding method can improve the signal-to-noise ratio and accurately detect the amplitude-frequency information as well as locate the mutation points of the transient signal.
维普
万方数据
