Abstract: During the last years, considerable efforts have been made for the optimization of numerical methods simulating the operation of the Circulating Fluidized Bed Combustors (CFBC), which imply both the accuracy increase and the computational cost decrease. A foremost goal is the efficient description of its operation under isothermal conditions and the in-depth understanding of the governing complex multiphase flow mechanisms. Grid construction and the calculation of the drag force experienced by the inert material in the two-phase flow are two important parameters for the optimization of the numerical approaches in the case of CFB simulation. The aim of the present investigation is to investigate both the effect of grid density distribution on simulation results and the validity of an anisotropic approach for the drag force calculation through an Energy Minimization Multi-Scale (EMMS) scheme. Grid density distribution is found to affect the numerical accuracy and the real time of simulations. Uniform grid density distribution is found to be the most efficient choice in terms of balance between computational cost and numerical accuracy. On the other hand, EMMS scheme improves the efficiency of detecting complex particle structures (clusters) without an explicit modeling of these spatio-temporal formations. Moreover, applying EMMS scheme only for calculation of the z-component of drag force yielded better numerical results compared to a constant interphase momentum exchange coefficient for all the three directions

Abstract: The arithmetic results from the formulation of an EMMS analysis for the calculation of drag coefficient between the co-existing phases in a CFB riser were implemented in a CFD code and three dimensional simulations of the isothermal flow of a 1.2 MWth CFBC unit were performed. Gas and inert material were modeled in an Eulerian fashion. Except from EMMS scheme, Gidaspow's correlation was also tested for reasons of comparison. Gidaspow's drag model is based on the assumption of homogeneous conditions inside a control volume, whilst the EMMS analysis encounters the effect of spatiotemporal multi-scale gas–particle structures on the induced drag force. Moreover, regarding the grid density, smaller control volumes enhance the validity of the homogeneous assumption. Thus, the effect of the grid density on the numerical results was also examined, using two uniform computational grids, consisting of hexahedral computational cells. Numerical results were compared with available experimental data, as far as the pressure drop along the bed is concerned. A good agreement with the experimental data was achieved in the case of the dense grid (43 mm/cell) using both approaches. In the case of the coarse grid (86 mm/cell), Gidaspow's correlation clearly under-predicted the experimentally measured pressure drop along the bed. This under-prediction was more significant in the lower part of the bed. On the other hand, the implementation of the EMMS scheme increased the accuracy of the model, mainly in the bottom region, since particles clustering was taken into account, a phenomenon which is more evident in latter region. In this area the drag force calculated via EMMS method is considerably less than the drag force calculated by Gidaspow's correlation. Overall, it is proven that the EMMS model is a very promising numerical tool for the more accurate drag force calculation since it reproduces numerically the effect of clustering mechanism on the time evolution of this complicated phenomenon, increasing the accuracy of the predictions without the need of denser numerical grids.

Abstract: Range and size increment of industrial applications regarding Circulating Fluidized Bed (CFB) technology raise numerous design and operating problems. Further insight into the governing and complex multiphase flow physics regarding CFB operation can be provided by CFD analysis. However, it has been proven during the recent years that multi-scale phenomena occurring in CFB reactors cannot be accurately reproduced by conventional models. Such models are unable to accurately predict the momentum exchange between the co-existing phases (gas and inert material) when calculating the drag coefficient. The main reason is that gas–solid two-phase flow features spatiotemporal multi-scale structures, named clusters, as the heterogeneous flow field is developing. Aiming to evaluate the particles tendency to aggregate in clusters and describe the effects of this complicated mechanism on the main characteristics of the induced multiphase flow field, the advanced Energy Minimization Multi-Scale (EMMS) analysis is applied. The EMMS scheme comprises of a set of equalities, constraints and a minimization energy equation, solved for a number of volume fraction's and slip velocity's values, using the optimization software General Algebraic Modeling System (GAMS).