Parameter estimation of a DC-DC converter using a Kalman Filter approach

This paper presents a system identification approach for power electronic systems, based upon a Kalman Filter (KF) approach. The proposed technique offers good parameter estimation accuracy, but with reduced mathematical complexity compared to other schemes, such as recursive least squares (RLS) based techniques. In this paper, the transfer function parameters of a dc-dc converter are estimated using both KF and RLS approaches. A brief assessment of both algorithms is presented, followed by a detailed simulation. Results demonstrate that the KF algorithm performs better than the RLS when an abrupt load change is applied to the buck converter; both in terms of parameter estimation accuracy and algorithm convergence time.