Application of the Real-Time Hilbert Huang Transform to a Noise Perturbed Buck Converter

Abstract

Extraneous noise signals can greatly downgrade the performance of a buck converter. In this paper, we investigate the use of an adaptive notch filter to dampen oscillations caused by perturbations of the input voltage. Typically, time-frequency analysis is carried out using traditional techniques such as the Fourier Transform or the Short-Term Fourier Transform. However, both of these methods require the signal to be stationary in its entirety or to be broken down into the sum of several stationary signals. Thus, these techniques are limited in their application. We propose using the Hilbert-Huang Transform to analyse in the timefrequency domain and identify the frequency of the non-stationary input. This can be used to tune the adaptive notch filter and the paper outlines a method for real-time application of the Hilbert Huang Transform.