Adaptive sensorless control for buck converter with constant power load

Abstract

The existing results mainly focus on the full-information control for converters with constant power load (CPL). More precisely, it is assumed that all of the states of the systems should be measured. There are no theoretical results on an adaptive sensorless control scheme for the converters feeding CPL with the guaranteed stability. Note that a sensorless control will facilitate the decrease of the overall cost and fault rate as well as the increase of the reliability of the systems. In this paper, the sensorless control problem for DC–DC buck converter with unknown CPL is addressed. The main contribution of this paper is to design a reduced-order generalized parameter estimation-based observer to simultaneously reconstruct the unmeasurable inductor current and unknown power load of the system. Borrowing the dynamic regressor extension and mixing technique, its main idea is to transform the state observation into the parameter estimation problem. Besides, the finite-time convergence of the observer can be ensured. It is noted that the buck converter is only regarded as an application example. As a matter of fact, this observer can be extended to a large class of converters with CPLs. Then, introducing the observed terms into an existing full-information controller, in a certainty equivalent manner, an adaptive sensorless control scheme is achieved. Finally, the performance of the designed controller is assessed via simulation and experimental results.