431824 A Modified Cohesion Model for CFD-DEM Simulations of Fluidization

Monday, November 9, 2015: 1:45 PM
254B (Salt Palace Convention Center)
Yile Gu, Ali Ozel and Sankaran Sundaresan, Department of Chemical and Biological Engineering, Princeton University, Princeton, NJ

Computational Fluid Dynamics (CFD) – Discrete Element Method (DEM) simulations have been used to study and explain the effects of cohesion on fluidization of particles (e. g. [1-4]). In order to accelerate simulations, all the studies mentioned above have decreased particle stiffness to allow for larger DEM time steps. The assumptions are that the flow patterns would not be altered when smaller particle stiffness is used. Indeed, it has been shown that the assumptions are valid for non-cohesive particles [5]. However, for cohesive particles, it has been clearly shown that the spring stiffness does affect flow behaviors in simulations of fluidization of cohesive particles [6]. Recently, Kobayashi et al. [7] proposed how one could modify the cohesion model so that the simulations with softer spring constants would yield similar results as those with realistic spring constants. The main limitation of the model by Kobayashi et al. is that their analysis does not account for the van der Waals force when particles are close to each other, but not in contact.

In this paper, we modified the cohesion model, such that the flow patterns would be insensitive (or only weakly sensitive) to spring stiffness even for cohesive particles. There are two notable differences between the model by Kobayashi et al. [7] and the one proposed here. Contrary to the study by Kobayashi et al. [7], the proposed modified cohesion model takes account of the van der Waals force even when particles are not in contact. In addition, the model proposed by Kobayashi et al. is only limited to linear spring model as contact model. We not only propose modifications to the van der Waals force model for linear spring model, but also for Hertzian contact model.

The proposed modified cohesion model has been tested by performing two-particle collision simulations and gas-fluidization simulations with Group A and Group C particles. For both cases, when the modified cohesion model is used, simulations yield nearly similar results independent of the value of the spring constant used in the simulations. In contrast, if the original cohesion model is used without modification, the predicted results depend significantly on the value of the spring constant.

[1] M. Ye, M. A. van der Hoef, and J. A. M. Kuipers, “A numerical study of fluidization behavior of Geldart A particles using a discrete particle model,” Powder Technol. 139 129-139 (2004).
[2] M. Ye, M. A. van der Hoef, and J. A. M. Kuipers, “The effects of particle and gas properties on the fluidization of Geldart A particles,” Chem. Eng. Sci. 60 4567-4580 (2005).
[3] Q. Hou, Z. Zhou, and A. Yu, “Micromechanical modeling and analysis of different flow regimes in gas fluidization,” Chem. Eng. Sci. 84, 449-468 (2012).
[4] J. E. Galvin and S. Benyahia, “The effect of cohesive forces on the fluidization of aeratable powders,” AIChE J. 60 473-484 (2014).
[5] Y. Tsuji, T. Kawaguchi, and T. Tanaka, “Discrete particle simulation of two-dimensional fluidized bed,” Powder Technol. 77, 79-87 (1993).
[6] R. Moreno-Atanasio, B. H. Xu, and M. Ghadiri, “Computer simulation of the effect of contact stiffness and adhesion on the fluidization behaviour of powders,” Chem. Eng. Sci. 62, 184-194 (2007).
[7] T. Kobayashi, T. Tanaka, N. Shimada, T. Kawaguchi, “DEM-CFD analysis of fluidization behavior of Geldart Group A particles using a dynamic adhesion force model,” Powder Technol. 248, 143-152 (2013).

Extended Abstract: File Not Uploaded