Coarse grid simulation of gas-solid flows in riser

Abstract

Gas-solid risers have been extensively used as multiphase reactors in circulating fluidized bed (CFB) system such as fluid catalytic cracking (FCC) and Circulating fluidized bed combustion (CFBC). In FCC, a riser facilitates cracking reactions which take place between fluidized catalyst and vaporised vacuum gas oil. Similarly in CFBC, combustion of coal particle occurs in the riser. The gas-solid flow hydrodynamics in riser plays a dominant role in governing the conversion of the chemical process, and in turn the performance of the CFB system. Therefore, several experimental and numerical studies have been conducted on the hydrodynamics of riser. Since experimental studies have been costlier and time consuming, detailed hydrodynamic modelling using computational fluid dynamics (CFD) has emerged as a promising alternative to investigate flows in riser.Two types of CFD models can be identified i.e. (i) Eulerian-Langrangian (EL) and (ii) Eulerian-Eulerian (EE) models. The Eulerian-Eulerian model considers both phases as interpenetrating continua. As a result, it allows a use of grid size larger than length scales of the flow structures. Therefore, this approach requires less computational power, and can handle large-scale geometry. Due to its computational edge, the Eulerian-Eulerian approach has been widely used to conduct the CFD simulations of gas-solid flows in riser. A literature review on the previous simulation studies revealed several shortcomings, and therefore, investigations on the impact of different modelling parameters were found to be necessary before applying the EE model to industrial-scale simulations. In this study, the EE gas-solid flow model has been used to carry out CFD simulations of risers. The effect of modelling parameters such as boundary conditions, drag models and solid phase closure models have been investigated. Finally, the EE model has been used to carry out industrial scale simulation of the lift engager to study the parametric effect of the lift gas velocity on the catalyst lift rate.The effect of inlet boundary conditions was investigated by using three different types of inlets for both gas and solid phases. The inlet arrangements had a profound effect on mixing patterns of two phases and energy balances in 2D riser. Furthermore, the effect of wall boundary conditions was also studied by considering the wall as a partial-slip or no-slip wall for the solid phase. A 3D full-scale riser was also simulated by implementing the boundary conditions similar to the experimental set-up. Although both 2D and 3D simulations could qualitatively predict the radial profiles of solid velocity and volume fraction, they could not predict even qualitative trend of the axial profile of solid volume fractions. Previous studies (Yang et al., 2004, Andrews Iv et al., 2005) showed that the conventional drag model such as that of Gidaspow’s could not capture the axial heterogeneity; however a use of the multiscale drag could remove this draw back.Thus, the drag derived using the energy minimization multi-scale approach was evaluated in another part of the study. The structure-based drag models could capture both axial and radial profiles of voidages, but only qualitatively. Furthermore, the use of different cluster diameter correlations with the EMMS framework gave more accurate predictions with reasonable qualitative agreement with the experimental data. Finally, the EE model was applied to study hydrodynamics of a complex industrial-scale lift engager. In this part, the effect of lift gas velocity on the catalyst lift rate and fluctuations therein were investigated to optimise operating gas velocity. Before conducting the parametric study, the effects of modeling parameters such as drag models and frictional pressure models on the hydrodynamics of the lift engager were also investigated.In summary, the study has been conducted by performing rigorous CFD simulations of both 2D and 3D risers. The effects of modelling aspects such as boundary conditions and drag models on the hydrodynamics predictions were investigated. The EE model with a structured-based drag from the EMMS model was found to be more effective in capturing an inherent heterogeneity of riser flows. However, the quantitative disagreements between the flow predictions and experimental data still persist.