Computationally Efficient Self-Tuning Controller for DC-DC Switch Mode Power Converters Based on Partial Update Kalman Filter

Abstract

In this paper, a partial update Kalman Filter (PUKF) is presented for the real-time parameter estimation of a DC-DC switch-mode power converter (SMPC). The proposed estimation algorithm is based on a novel combination between the classical Kalman filter and an M-Max partial adaptive filtering technique. The proposed PUKF offers a significant reduction in computational effort compared to the conventional implementation of the Kalman Filter (KF), with 50% less arithmetic operations. Furthermore, the PUKF retains comparable overall performance to the classical KF. To demonstrate an efficient and cost effective explicit self-tuning controller, the proposed estimation algorithm (PUKF) is embedded with a Bányász/Keviczky PID controller to generate a new computationally light self-tuning controller. Experimental and simulation results clearly show the superior dynamic performance of the explicit self-tuning control system compared to a conventional pole placement design based on a pre-calculated average model.