CFD-DEM simulation of fluidized bed with an immersed tube using a coarse-grain model

Highlights

• A new coarse-grain (CG) model for CFD-DEM simulation is proposed.

• The CG CFD-DEM is used to investigate the fluidized bed with an immersed tube.

• The applicability and accuracy of the CG CFD-DEM is verified.

• The CG CFD-DEM can reasonably predict the erosion of the immersed tube.

• The computational efficiency is significantly improved due to the CG model.

Abstract

The computational fluid dynamic combining the discrete element method (CFD-DEM) is widely used to comprehend the mechanism of gas–solid flow. However, as a Lagrangian method, DEM suffers from the huge cost for tracking each particle in the systems when applied to industrial applications with billions of fine particles, regardless of the great advances in computing capacity of computers. To address this issue, a coarse-grain (CG) model is proposed in this paper, where the original fine particles are represented by a small number of large-sized CG particles. In the current study, the new-developed CG model combining CFD-DEM (CG CFD-DEM) is applied to investigate the fluidized bed with an immersed tube. The applicability and accuracy of the proposed CG CFD-DEM are validated by the good agreement of the macroscopic behaviors predicted by simulations and those obtained from experiments. For the first time, the CG CFD-DEM is used to predict the erosion around the tube based on the shear impact energy model (SIEM), and the computational results calculated by CG CFD-DEM and original CFD-DEM agree well, which demonstrates that the proposed model can predict the erosion of the tube in the fluidized bed accurately and efficiently. Besides, as expected, the calculation time is drastically reduced due to the CG model, which makes the simulations of large-scale industrial applications accessible.

Introduction

Fluidized beds are frequently used in various industries due to their excellent heat and mass exchange efficiency (Asegehegn et al., 2011b, Goldschmidt et al., 2004, Liu et al., 2020a, Taghipour et al., 2005, Tsuji et al., 2008, Yue et al., 2019, Zhang et al., 2019). In many applications, immersed tubes are introduced into the fluidized beds to enhance the transfer and conversion rates, control the bed temperature and facilitate the particle mixing (Asegehegn et al., 2011a, Hou et al., 2016, Li et al., 2011, Schreiber et al., 2011, Wahyudi et al., 2016, Yurong et al., 2004). However, understanding the gas–solid flow properties is difficult ascribed to the complex underlying mechanism in the systems and shortcomings in research methods, which greatly impede their design and scale-up. The presence of tubes further complicates the gas-flow dynamics of the systems and then increases the difficulty of investigation (Asegehegn et al., 2011b). Therefore, an in-depth study of gas–solid behavior in the fluidized bed with immersed tubes has aroused the interest of researchers. In recent years, many experiments have been conducted to gain further insight into the various characteristics of fluidized bed with immersed tubes involving heat transfer (Kamble et al., 2014), bubble behavior (Asegehegn et al., 2011a, Asegehegn et al., 2011b), and erosion around the tube surface (Wiman and Almstedt, 1997). However, experimental measurements have some limitations, e.g., the harsh operational conditions, high expense, and restricted optical access (Ostermeier et al., 2019, Qi et al., 2019). Thanks to the rapid progress in available computational resources and numerical models, numerical simulation gains particular importance as a powerful tool for the prediction of gas–solid flow (Almuttahar and Taghipour, 2008, Liu et al., 2020b), e.g., CFD-DEM (computational fluid dynamic - discrete element method) and CFD-DPM (computational fluid dynamic - discrete phase method). As typical Eulerian-Lagrangian methods, CFD-DEM can track every particle individually taking into account the particle–particle and particle–wall interaction, and it has been widely used in the simulation of fluidized beds (Gan et al., 2016, Lu et al., 2014, Moliner et al., 2019, Su et al., 2011, Zhao et al., 2017); while CFD-DPM neglects the interaction between particles, and it is suitable for the simulation of dilute flow. To date, the industrial application of CFD-DEM with billions of particles is still limited because of excessive calculation cost even though great advances have been made in the computing capacity of computers. The calculation time, therefore, becomes extremely long to the extent that the simulation will last for months by a single personal computer. Hence, coarse-grain model is proposed to overcome exhaustive computational cost existing in the present DEM method because it can notably reduce the number of particles tracked in the numerical domain. The coarse-grain model combining the CFD-DEM method (CG CFD-DEM) has been proved to be effective by many researchers and applied to a variety of situations. Sakai and coworkers simulated the macroscopic behavior in three-dimensional fluidized beds numerically and experimentally using the CG model (Sakai et al., 2014). They simulated a group of original particles with large-sized particles, referred to as the CG particles. In their CG model, the total energy was estimated to agree between the CG particle system and the original particle system. They verified the effectiveness of the CG model through the test simulation, where the calculated particle size was the same as the CG particle without using the CG model. Chu et al. developed a comprehensive CG CFD-DEM to simulate multiphase flow in a cyclone, and they analyzed the error caused by this model (Chu et al., 2016). In their model, original particles with the same properties, e.g., size and density, can be grouped into a CG particle. They emphasized that particles included in the CG particle had equal translational and rotational movement. The conclusion demonstrated that the CG CFD-DEM model was more accurate in the scenario where the flow was dilute. Lu et al. extended the CG method to the simulation of liquid–solid reacting flows (Lu et al., 2016). They determined the diameter of the CG particle with the energy minimization multiscale (EMMS) model and calculated the collision force using the collision diameter. Based on CG CFD-DEM, Lan et al., 2020, Lu et al., 2020 studied the residence time distribution of particles. Hu et al. (2019) investigated the effect of different operating parameters on the coal gasification process with CG model. About the coarse ratio, Mu et al. (2020) emphasized that it possessed an upper limit which was related to the length scales at the mesoscale that can still be resolved.

Wear occurs when a surface is subjected to the intensive impingement of particles for being exposed to the dense gas–solid flow (Finnie, 1960). This will notably shorten the service life of the device and affect the continuous operation and the economics of fluidized bed plants. Therefore, researches on the wear mechanism should be paid attention to. Investigating erosion properties with numerical methods has become a focus in this field, and many erosion models have been proposed such as the kinetic theory erosion model and kinetic energy dissipation erosion model based on the Eulerian-Eulerian method. These models have low demand for computing resources as opposed to the Eulerian-Lagrangian models whose detailed information of particles is obtained by the Lagrangian method. However, due to the latent defect of those methods that only mean values on a small computational volume can be given, developing a new erosion model to accurately predict the wear of the immersed tubes is an urgent need. Zhao et al. (2017) first proposed a quantitative erosion model called shear impact energy model (SIEM) based on Ashrafizadeh’s work (Ashrafizadeh and Ashrafizadeh, 2012) and proved its accuracy through comparing the simulation results with previous experimental data. In the current work, SIEM is also adopted to study the erosion of the immersed tube in the fluidized bed.

Up to now, the CG model is seldom applied to the fluidized bed with an immersed tube, especially the wear occurring on it. In this study, a CG CFD-DEM is proposed and verified to be applicable and accurate to investigate the fluidized bed with an immersed tube through comparing the macroscopic results obtained from CG CFD-DEM and that predicted by original CFD-DEM as well as experiment results, i.e., the flow pattern, the solid velocity, the pressure drop, the bed height, and the void fraction around the tube. In the CG model developed by us, the energy dissipation rate of the CG system agrees with that of the original system through modifying key collision parameters, referred to as the coefficient of the sliding friction. For the first time, the CG CFD-DEM combining SIEM is applied to the prediction of erosion around the tube. Besides, the simulation time consumed by CG CFD-DEM and original CFD-DEM for a given time size is recorded to prove the capacity of reducing computing cost for the CG model.