Performance Comparison of Model Predictive Control Methods for Active Front End Rectifiers

Active front-end rectifiers are tasked with generating high-quality, low-distortion sinusoidal line currents in the presence of adverse circuit conditions. When suitable control methods are applied to such rectifiers, significant performance improvements can be realized, especially under abnormal utility conditions. Although several control methods based on model predictive control platforms have been recently developed, there is a lack of comparative studies of these methods in the literature. In this paper, the details and theoretical background of four model predictive control methods, namely, current control, virtual flux control, direct power control, and virtual flux direct power control, are presented. Then, the performance of these methods was compared by the way of experiments to determine the respective quality of the line current under various conditions, such as unbalanced input voltages, input voltage distortion, uncertainty in the parameters, and dc voltage fluctuation. In summary, based on the results of the experiments, the virtual flux control scheme was found to be distinctly superior to the other three schemes for distorted and unbalanced line voltages.