A particle–grid method for Euler–Lagrange approach

• Hydrodynamics of a cold-flow spouted fluidized bed have been investigated.
• Physical values of solid phase are computed using additional particle–grid.
• Application of particle–grid method allows the refinement of fluid grid resolution beyond particle size.
• Stochastic Euler–Lagrange and DEM model can reproduce the right fluidization regimes.
• Stochastic Euler–Lagrange model reaches, however, earlier the maximum bed expansion.

In Euler–Lagrange approach, the physical values of fluid and solid phases are computed in a common grid, namely fluid grid. The refinement of the fluid grid resolution beyond the particle scale leads to inconsistency with the basis of using extended Navier–Stokes equations for gas–solid flow (Anderson et al., 1967). In this study, an additional particle–grid is applied, in which the physical values of solid phase are computed. To investigate the influence of the particle–grid application on the simulation accuracy, the numerical results obtained by Euler–Lagrange approach combined with a deterministic collision model (known also as Discrete Element Method (DEM)) are validated with measurements obtained from a lab-scale spouted fluidized bed. The results confirm that the particle–grid method allows the variation of the fluid grid resolution independent of the particle size and consequently improves the calculation accuracy. In the second part of this work, the simulation results obtained from the extended Euler–Lagrange/DEM model are compared with the simulation results obtained from the Euler–Lagrange approach combined with a stochastic collision model. Two different fluidization mass flow rates are considered to analyse the ability of the used simulation approaches to predict the hydrodynamic behaviour of the gas spouted fluidized bed. The results show that both techniques can reproduce the right fluidization regimes including the bubble size and the bed expansion. Deviations from the experimental data in the jet zone and during the final stage of the bubble formation are, however, observed. The reasons of these discrepancies in predicting the dynamic behaviour of the bed, the advantages and limitations of the two approaches are demonstrated.

Figure optionsDownload as PowerPoint slide