UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Numerical simulation of particle separation in centrifugal air classifiers Barimani, Mohammad


The demand for fine mineral powder in various industries has stimulated creative methods for separating the fine portion of particles from a mixture. Among the many different types of classifiers invented, centrifugal rotor air classifiers are characterized by their capability in producing ultra-fine products with a cut-size as low as 3um. Classification occurs due to the size-dependence of aerodynamic and inertial forces acting on particles: coarse particles have a higher ratio of centrifugal force to aerodynamic drag than do fine particles, and therefore are preferentially ejected to the classifier perimeter. Therefore, the high speed rotor located inside the classifier is key to classification. Computational fluid dynamics (CFD) is utilized in this study to investigate the motion of calcium carbonate particles in a rotor classifier. The single phase flow in two- and three-dimensional models of the rotor is computed. Two turbulence models, namely K-Omega and RSM, are applied to close the Reynolds-averaged Navier-Stokes equations. Once the single phase flow has been computed the motion of solid particles is simulated using the Discrete Phase Model. This model ignores particle-particle interactions and the influence of the particles on the air flow. The motion of the particles is coupled to a statistical model of the turbulent velocity fluctuations. By tracking hundreds of particles, the efficiency for a variety of hypothetical classifiers is estimated. Though the CFD models, in comparison with experiments, cannot accurately predict the absolute cut-size values, they have proved effective in predicting cut-size shifts as a result of rotor geometry modification or alternative operating conditions. Based on these simulations two new rotors were built and the change in cut-size was predicted within 30% accuracy. Based on the paths of a large number of particles tracked in various operating conditions, regions in the rotor with very high particulate concentrations are identified. We speculate that this elevated concentration makes particle-particle interactions much more important than would be expected based on the feed concentration, which could in turn reduce the acceptance of the smallest particles.

Item Citations and Data


Attribution-NonCommercial-NoDerivs 2.5 Canada