The study of spectrum estimation in the power electronic devices based on wavelet analysis

The EMI spectrum plays an important role in the EMC prediction of the power electronic devices. A spectrum estimation method is presented, which combines the good ability of Fourier analysis in accurate frequency localization and the features of wavelet analysis both in saltation signal processing and in adaptive signal processing, to apply a wavelet thresholding to the EMI characteristic of the system. This thresholding method has been verified by simulation and the experiment results in a power converter.