Design and analysis of a novel adaptive learning control scheme for performance promotion of grid-connected PV systems

Abstract

This paper addresses a hybrid adaptive iterative learning control strategy for controlling power converters that are used in photovoltaic systems to enhance maximum power point tracking capability in the presence of variable atmospheric conditions. The adaptation of the controller to the fast-changing environmental conditions is provided by a fractional-order proportional-integral type learning control mechanism. The developed control scheme is integrated into a grid-connected current-source flyback inverter to highlight the improvements in performance criteria such as convergence speed during transients, tracking accuracy, steady-state oscillations, and robustness. The performance analyses are carried out under various scenarios. The obtained results reveal that the dynamic response of the system is considerably increased under erratic atmospheric conditions while steady-state oscillations are decreased for stable operation conditions. The maximum absolute error that indicates the robustness of the proposed controller is decreased from 2.3704 to 2.1920. In addition, the error deviations of the proposed control algorithm are below 10%. The variance of the error, which shows steady-state stability, is reduced from 2.5123 to 1.6152. Also, the proposed controller reduces the amount of control energy by 20% when compared to the PI controller. Furthermore, the values of IAE and ISE are reported 10% lower in the proposed controller.