Large Eddy Simulation of particle-laden swirling flow with a presumed function method of moments

Abstract

In the present paper an Eulerian–Eulerian polydisperse two–phase flow model for particle–laden gas flows is presented and has been validated against experimental data using Large Eddy Simulation (LES). Modelling polydispersity of particle dynamics is based on a moments model, which considers the size dependency of particle motion and phase interaction forces by transporting the moments of the particle size distribution function with their individual moment transport velocities. LES filtering was carried out using a weighted filtering approach in analogy to the density weighted Favre averaging. The filtered gas phase equations were closed using a standard subgrid–scale model and the interaction of subgrid–scale fluctuations with the particles has been neglected. Closure with a Beta–based presumed distribution, geometry of moment space and validity of moment sets are discussed. The overall computational cost for the polydisperse LES is only about twice those of the single–phase one, which is very favourable. Comparison of the numerical results with the experimental data of Sommerfeld and Qiu (1991) [Sommerfeld, M. and Qiu, H.H. (1991) 'Detailed measurements in a swirling particulate two–phase flow by a Phase–Doppler Anemometer', Int. J. Heat and Fluid Flow, Vol. 12, No. 1.] shows that complex two–phase flows can be predicted accurately.