Application of a hybrid Lattice Boltzmann–Finite Volume turbulence model to cyclone short-cut flow

Abstract

This paper introduces a hybrid turbulence model for studying the phenomenon of cyclone short-cut flow: While the global flow situation comprising inflow, cyclone flow, and outflow is described by a Finite Volume based Reynolds-Averaged Navier–Stokes (RANS) turbulence model, the top annulus region of the cyclone is modelled by an embedded fine-grid Lattice Boltzmann based Large Eddy Simulation (LES). Our hybrid turbulence model combines the numerical efficiency of the Lattice Boltzmann algorithm on an equidistant lattice with the flexibility of Finite Volume simulations on arbitrary grids. From an algorithmic point of view, the embedded Lattice Boltzmann simulation receives information from the Finite Volume simulation via an unsteady corona boundary condition. The flow pattern in the Finite Volume simulation is, in turn, locally influenced by the underlying Lattice Boltzmann simulation via a direct forcing method. Unlike standard RANS simulations that predict a dominant short-cut flow to the entry of the vortex finder, our hybrid turbulence model predicts that the short-cut flow is almost completely dispersed into the rotating main flow by unsteady turbulent eddies. A numerical experiment in which solid particles were introduced at the top wall indicates that the short-cut flow does not reduce separation efficiency as predicted by RANS simulations. We nest a Lattice Boltzmann based LES into a Finite Volume based RANS simulation in order to resolve the short cut flow phenomenon in gas–solid cyclones. This hybrid turbulence model is computationally very efficient and results agree well with measurements. ► We nest a Lattice Boltzmann based LES into a Finite Volume based RANS simulation. ► The hybrid turbulence model is applied to a gas-solid cyclone. ► The hybrid turbulence model is computationally efficient and accurate.