Airflow-assisted separation of tobacco lamina and light impurities based on CFD–DEM coupling and experimental validation

Air separation is widely used for removing light impurities from tobacco lamina, but the separation performance is strongly affected by airflow organization, particle morphology, and feeding intensity. In this study, a two-way coupled CFD–discrete element method (DEM) model was established to investigate the airflow-assisted separation of tobacco lamina and light impurities in a winnowing chamber. Tobacco lamina were modeled as rigid non-spherical particles, while plastic-film fragments and jute fibers were represented using flexible particle models. The results showed that the chamber flow was composed of a main transport path, a top suction path, and several recirculation zones. The sweeping-converging region beneath the top cover promoted light-impurity capture, whereas the central recirculation zone increased particle retention and re-entrainment. The optimal operating condition was obtained at an upper inlet velocity of 3.3 m/s, a lower inlet velocity of 5.3 m/s, and a solid-phase mass flow rate of 9220 g/h. Under this condition, the simulated impurity removal efficiency, tobacco loss rate, and comprehensive separation index were 96.0%, 1.0%, and 95.04%, respectively. Experimental validation showed good agreement with the simulation, with the relative error of the comprehensive separation index below 3.5%. The results provide guidance for airflow regulation and structural optimization of tobacco impurity-removal equipment.