In this study the multiphase particle-in-cell (MP-PIC) method is used for the simulation of two pilot-scale circulating fluidized beds (CFBs) with quartz sand belonging to Geldart’s group B as bed material. The simulations were performed using a homogeneous drag model as well as a structure dependent drag model based on the energy minimization multi-scale method (EMMS). The results are compared with experimental data from literature as well as experiments. The simulations with the EMMS based drag model show a good agreement of the time-averaged axial solids concentration, circulation rate and riser pressure drop. Furthermore, a lower grid sensitivity is observed compared to the homogeneous drag model. In contrast to the conventional drag model a dense bottom zone is predicted by the EMMS based drag model. An overprediction of the solid concentration in the dense bottom zone is presumably due to an overprediction of the cluster diameter that is calculated using an empirical cluster diameter correlation. This shows the necessity for a new meso-scale cluster correlation for the simulation of Geldart B particles. Furthermore, the results of the time-averaged radial solids concentration differ from the expectations of a core-annulus flow indicating that a mesh refinement at the walls is necessary. Finally, the importance of using a realistic particle size distribution is identified.

在这项研究中，多相颗粒胞内 (MP-PIC) 方法用于模拟两个中试规模的循环流化床 (CFB)，其中以 Geldart 的 B 组石英砂为床材料。使用均质阻力模型以及基于能量最小化多尺度方法 (EMMS) 的结构相关阻力模型进行模拟。结果与来自文献和实验的实验数据进行了比较。使用基于 EMMS 的阻力模型的模拟显示时均轴向固体浓度、循环速率和立管压降的良好一致性。此外，与均质阻力模型相比，观察到较低的网格灵敏度。与传统的阻力模型相比，基于 EMMS 的阻力模型预测了密集的底部区域。密集底部区域中固体浓度的过度预测可能是由于使用经验簇直径相关性计算的簇直径的过度预测。这表明对于 Geldart B 粒子的模拟需要新的中尺度簇相关性。此外，时间平均径向固体浓度的结果与岩心环空流动的预期不同，表明壁面的网格细化是必要的。最后，确定了使用实际粒度分布的重要性。此外，时间平均径向固体浓度的结果与岩心环空流动的预期不同，表明壁面的网格细化是必要的。最后，确定了使用实际粒度分布的重要性。此外，时间平均径向固体浓度的结果与岩心环空流动的预期不同，表明壁面的网格细化是必要的。最后，确定了使用实际粒度分布的重要性。