Frequency-domain identification based on pseudorandom sequences in analysis and control of dc power distribution systems: A review

Abstract

Frequency-domain identification based on wideband techniques has become a popular method in the analysis and control of various dc power distribution systems. In the method, a single converter or a system is perturbed by an external wideband voltage or current injection, the resulting voltage or current responses are measured, and Fourier analysis is applied to extract the spectral information of the measured variables. Most often, the system or converter input and output impedances and the loop gain are the quantities of interest. One class of perturbation signals, pseudorandom binary sequences, has become widely used because most of such signals can be generated using simple shift-register circuitry. As the signals are deterministic and binary, they are well suited to perform measurements on power-converter systems in real time, allowing fast response to system variations through, for example, adaptive controllers. This article reviews the pseudorandom binary sequences applied to dc power distribution systems and discusses their advantages as well as limitations. The conventional maximum-length binary sequence, inverse-repeat binary sequence, discrete-interval binary sequence, and orthogonal binary sequences are considered. Several experimental results from various dc power systems are presented and used to demonstrate the applicability of the discussed methods.