Skip to Main Content
This paper investigates predictive average current control schemes of a bi-directional DC-DC converter. By properly arranging the sampling, duty cycle updating, and controlling points, three predictive control methods are derived for better steady-state and dynamic performance. The less the delay between the sampling point and controlling point has, the better the stead state and dynamic performance is. System stability with the three predictive current control methods are analyzed in the z-domain considering inductance variation and it turns out that the control methods are stable as long as the inductance value used in the controller is less than twice of the real inductance value of the converter. The effect of inaccurate model and parameters on static current error is investigated and an additional integral controller with small gain is proposed for eliminate such steady-state error. Experimental test results from a prototype converter rated at 360 W/ 10 A within a DC micro gird validate the excellent performance of the proposed schemes.