CFD-DEM Model Research Articles

Comprehensive multivariate sensitivity analysis of CFD-DEM simulations: Critical model parameters and their impact on fluidization hydrodynamics

The development of CFD-DEM is critical for investigating particle phenomena and their coupling with reactor transport. However, there continues to be considerable uncertainty in the selection of model parameters because of limitations in: (a) experimental measurements of multi-particle interactions, and (b) computational resources which have restricted most numerical studies to 2D simulations, in very small-scale systems (<50k particles) and/or to local sensitivity analysis. The focus of this study is to identify critical model parameters in 3D CFD-DEM simulations of fluidized beds through multivariate sensitivity analysis and quantify their impact on hydrodynamics. Towards this end, thirteen model parameters are considered and the sampling design matrix is constructed using the Morris-One-At-a-Time (MOAT) screening method. 3D CFD-DEM simulations with almost 170,000 glass bead particles (0.4 mm diameter) are conducted in a small rectangular pulsating fluidized bed, selected because of its repeatable bubbling patterns. Detailed bubble and particle dynamics data from 250+ simulations show that: (a) choosing exceedingly low normal spring stiffness has strong implications on particle velocities; (b) the impact of all contact dissipation parameters (normal restitution, friction and tangential damping) is tightly coupled and sensitivity to any one hinges on the choices of others; and (c) the stability of bubble patterns is contingent on their choices and almost-ideal as well as extremely dissipative systems exhibit no patterns. In addition, by investigating particle dynamics inside and around bubbles, we derive a working expression for the optimal choice of spring stiffness. Overall, this first-of-its-kind analysis provides important guidelines for CFD-DEM model parameter selection and the statistical framework developed here provides a robust strategy for the fundamental investigation of other particle-scale phenomena and simulation-based reactor design and optimization.CFD-DEM, gas-solid fluidization, pulsating reactor, multivariate sensitivity, linear spring-dashpot model, spring stiffness

CFD-DEM investigation of the fluidization of binary mixtures containing rod-like particles and spherical particles in a fluidized bed

In the energy conversion process, the inert bed material such as sand with spherical shape is usually added to the fluidized bed for the proper fluidization of the non-spherical biomass particles. Understanding fluidization of the binary mixtures containing the biomass and the bed material is therefore significant for better utilization of the biomass. To this end, CFD-DEM model is used to simulate the fluidization of the binary mixtures containing the rod-like particles that represent the light and large biomass particles and the spherical particles that represent the dense and small bed material, in which all particles are described by super-ellipsoid model. Five granular systems with different volume fractions of the rod-like particles but the same particle volume are simulated in this paper. The simulation results demonstrate that the minimum fluidization velocity would increase with the volume fraction of the rod-like particles. The addition of the spherical particles is beneficial to the fluidization of the rod-like particles. The mixing rate in binary granular system and the distribution of the coordinate number would be affected by the gas superficial velocity and the volume fraction of the rod-like particles. For the orientation of the rod-like particles, its distribution in binary granular system is different from that in monodisperse rod-like particle system. But whether in binary granular system or monodisperse rod-like particle system, there is apparent wall effect on the particle orientation.

A semi-resolved CFD-DEM model for seepage-induced fine particle migration in gap-graded soils

A semi-resolved CFD-DEM model is proposed to model the seepage-induced fine particle migration within the gap-graded soils consisting of fine and coarse particles. In this model, a resolved method, i.e., Fictitious Domain (FD) method is used to resolve the fluid flow around the coarse particles and the un-resolved approach based on the locally averaging theory is adopted to describe the interactions between the fine particles and the fluid flow for the computational efficiency. The semi-resolved CFD-DEM model is validated using several simple examples and is further used to analyze the fine particle migration and associated piping erosion in gap-graded soils.

Modeling analysis of argon gas flow rate’s effect on pre-mixed powder separation in laser metal deposition process and experimental validation

Pre-mixed powder is frequently used powder supply to fabricate functional gradient material by laser metal deposition (LMD). Argon gas flow blows the powder mixture following feeding pipe to melt pool in the LMD process. The powder mixture easily separates since the ingredient particles have different accelerations which are caused by different densities and sizes under the dynamic interaction with argon gas flow. This study investigated the argon gas flow rate’s effect on pre-mixed powder separation using modeling methodology. A CFD-DEM model, which gave consideration to both particle-particle collision and particle-argon gas flow interaction, was employed to simulate the pre-mixed powder flow in the feeding pipe. Three argon gas flow rates 6, 7, and 8 m/s were selected, analyzed, and compared based on their effects on powder mixture separations. Pre-mixed Cu and 4047 Al powders with equal volume percentages (50 to 50%) were investigated during their transportation process under three argon gas flow rates. All particles’ dynamic flow behaviors and the powder distributions were simulated and observed. The volume percentage of each kind of powder was plotted by quantifying the distribution of different particles after exiting the nozzle. It can be found that the intersection point of both powder volume percentages appeared increasingly earlier along with the increasing argon gas flow rate. To prove the correctness of the simulation results, the CFD-DEM model in this study was then validated by an experiment done by authors before. The results from this study are valuable contributions to the research of functionally graded material fabrication with pre-mixed powder through LMD process.

Superheated steam drying of single wood particles: A characteristic drying curve model deduced from continuum model simulations and assessed by experiments

ABSTRACTComputational fluid dynamics (CFD) and CFD coupled with discrete element method (DEM) are powerful approaches to describe the drying behavior of real drying towers. In these approaches, the heat and mass interaction between fluid phase and the wet solid is required as an essential input. In this work, a new methodology for establishing simple drying model of single wood particles is presented. First, a spatially resolved continuum-scale model that describes the coupled heat and mass transfer within a single porous wood particle during superheated steam drying under atmospheric pressure is developed. The thermophysical properties of wood particle required as input data for the continuum model are determined experimentally. Then sophisticated continuum model is reduced to a simpler lumped model, which is referred to as the characteristic drying curve (CDC) model. The continuum model simulations are performed in a board range of operating drying conditions and the associated results are used to establish correlations for the CDC model parameters. The results of both sophisticated and reduced models are validated against the experimental observations made using a magnetic suspension balance. The sensitivity analysis performed with the continuum model indicates a strong dependency of the critical moisture content on particle size and drying condition. The established CDC model can be implemented in the CFD or CFD-DEM model of superheated steam dryers.

Numerical Studies of Separation Performance of Knelson Concentrator for Beneficiation of Fine Coal

ABSTRACT The separation of fine coal in a laboratory Knelson concentrator was studied through experimental and numerical simulation. The flow field characterized of Kenlson concentrator simulated in a three-dimension laboratory Knelson concentrator by using CFD modeling. The results show that the flow regime can be treated as laminar flow, which propitious to layering for particles. In the riffle part, eddies were created by the rotational flow and fluidizing water. The flow of fluids changes from turbulence to laminar, which was propitious to separation for particles. So, the separation processes were realized in riffle part. The CFD-DEM modeling was applied for the first time in the study of the influence of centrifugal force and fluidizing water rates on the separation performance of a laboratory Knelson concentrator. The simulation results indicate that the Ep value was increased with decreasing the centrifugal force at the constant fluidization water rates, and increased with increasing the fluidization water rates at the constant centrifugal force. The simulation results were consistent with the experimental results and the CFE-DEM model was effective and reliable.

CFD-DEM modelling of circulation frequencies and residence times in a prismatic spouted bed

Coupled CFD-DEM simulations of a three-dimensional prismatic spouted bed with liquid injection were performed. Spouting stability is investigated by means of pressure drop fluctuation analysis and high speed camera recording. To determine the influence of flow stability on coating quality, droplets are injected in a post-processing step and deposited on colliding solid parcels. In order to estimate liquid holdup in the bed, heat transfer and subsequent evaporation of the surface liquid film were modelled. The coating quality of different process conditions is compared for a simulation time of 10s and found to be broader distributed in the dense stable spouting operating regime. The analysis of particle circulation frequencies shows slower circulation under stable spouting conditions, resulting in a slower coating progress. This explains the inhomogeneous coating quality in the first 10s for these cases. Nevertheless, the frequencies are less broadly distributed compared to the instable regime, which is advantageous for obtaining a homogeneous layer in real-time coating processes.

CFD-DEM modeling of strongly polydisperse particulate systems

Particulate systems can be readily modeled and simulated with the discrete element method (DEM). Large ranges in grain sizes, however, pose a serious challenge: The number of particles per volume increases significantly and computational time steps need to be reduced. Especially large systems require an efficient treatment of the smallest particles present. We derive and discuss models for two limiting cases. (a) For clearly separated size distributions, we consider the smaller particles as continuous fields. This corresponds, for example, to fines transport through a coarse granular bed including their deposition and release and is validated against measurements on such a system (approx. 105 particles, fine-to-particle-diameter ratio 0.105, up to 200% increase in pressure drop). (b) For less strongly separated distributions, e.g. one broad range of sizes, we introduce effective parcels consisting of sub-scale particles with various diameters. This model is applied to numerical experiments with a horizontal gas jet penetrating packed beds with bi- and monodisperse compositions and forming cavities as found in blast furnaces. The results show a good agreement with what one would expect from the full size distribution at fractions of the actual particle number (Nbi ≈ 3 ⋅ 105, Neff ≈ 7 ⋅ 104). Finally, we discuss how these models might be generalized for the application to a wide range of particulate systems.

Dynamic modelling of fluidisation in gas-solid bubbling fluidised beds

Formation of fluidised bubbles in a fluidised bed has significant impact on the design and scale-up of the bubbling fluidised bed. To fully understand bubble motions in the fluidised bed, fluidised bubble dynamics in a gas-solid fluidised bed is investigated using CFD-DEM numerical simulation. To reveal the mechanism of bubble formation in the bubbling fluidised bed, the time autocorrelation approach is adopted to identify bubble occurrence at different bed height positions. A non-dimensional correlation accounting for the effect of local velocity field on the bubble size, which is obtained by modifying the Darton's model, is proposed for predicting the bubble size at given bed height positions. The non-dimensional correlation for predicting the size of the fluidised bubbles at various given bed height positions is obtained by regression based on the CFD-DEM modelling results. The simulation results indicate that small bubbles with high passing frequency dominate in the bottom region of the fluidised bed, while large bubbles are formed in the zone close to the free surface of the fluidised bed. The probability of the appearance of large bubbles increases as the height measured from the bottom of the fluidised bed increases, which may be attributed to the coalescence of small bubbles; this is also reasonably consistent with the results predicted using the Darton's model and experimental observations. It was observed that the sizes of the fluidised bubbles at the given height positions correlate with the local dynamic parameters characterised by the Reynolds number and the dimensionless height ratio.

CFD-DEM modeling on air impact densification of equal spheres: Structure evolution, dynamics, and mechanism

This paper presents a numerical modeling on the packing densification of uniform spheres under air impact using a combined Computational Fluid Dynamics (CFD) and Discrete Element Method (DEM). In the whole process, the packing structure evolution, corresponding dynamics and densification mechanism are comprehensively studied. Macro and micro properties such as packing density, coordination number, radial distribution function, as well as forces in the packings at different stages are characterized and compared. The results show that air impact can realize the transition of particle packing from random loose to random close state at appropriate conditions. In this duration, the depth-averaged normal force increases linearly with the height. Meanwhile, numerous normal forces are close to the horizontal at the final packing stage due to the effect of air impact, and their distributions indicate exponential decay law. While the distribution of the fluid-particle interaction forces increases with the height. Local packing structure evolution demonstrates that the dominating densification mechanism under the air impact is ‘pushing filling’.

Hydrogen production by biomass gasification in a supercritical water fluidized bed reactor: A CFD-DEM study

In this study, we used CFD-DEM model to simulate glucose gasification in a supercritical water fluidized bed reactor (SCWFBR). Simplified reaction kinetic model, including glucose decomposition, water gas shift reaction and methanation reaction, was used. Particle distributions, gas composition and reaction rate profiles were analyzed. Moreover, the influence of chemical reaction on flow behavior, and the effects of operation parameters, such as wall temperature, flow rates and initial bed height were investigated. The results show that high wall temperature, low flow rate and high initial bed height are in favor of gasification. However, lower H2 yield, gasification efficiency (GE) are found at flow rate below the minimum fluidization velocity. The simulation results show good agreement with the experimental results, demonstrating the validity of this model. The information obtained by the numerical study is useful for the operation and optimization of the gasification reactor.

Understand solids loading effects in a dense medium cyclone: Effect of particle size by a CFD-DEM method

Industrial cyclones, such as gas, hydro and dense medium, are widely used in chemical, mineral and process industries to separate particles from fluids or classify particles by size, density or other solid properties. It is well known that the loading of particles can significantly affect fluid flow in such cyclones but the specific effect can be confusing due to a limited fundamental understanding of the working mechanisms involved. In this work, the problem is partially tackled by analysing the influence of particles of different size on the medium flow in a dense medium cyclone (DMC) by using a combined approach of Computational Fluid Dynamics (CFD) and Discrete Element Method (DEM) (CFD-DEM). In the CFD-DEM model, CFD is used to model the motion of slurry medium by numerically solving the local-averaged Navier-Stokes equations at a computational cell scale, facilitated by the Volume of Fluid (VOF) and Mixture multiphase models. DEM is then used to model the motion of discrete particles by applying Newton's laws of motion, facilitated by a coarse-grained (CG) method to model small coal particles. It is found that both the particle and medium flows are sensitive to the particle size. The most notable finding is that spatial distributions of solid flow patterns form due to particles of different size have different trajectories in the cyclone. These lead to significant different distribution of volumetric particle-fluid interaction forces which cause different effects on the fluid flow. The findings are useful for developing a better understanding the working mechanisms of solids loading effects in DMCs and also in other similar swirling multiphase flow systems.

CFD-DEM simulations of early turbulent solid–liquid mixing: Prediction of suspension curve and just-suspended speed

Solid–liquid mixing as a unit operation still faces considerable challenges, notably regarding the prediction of the impeller speed required to suspend the particles (Njs), the fraction of suspended solids and the homogeneity of the suspension at a given speed. In this work, we extend to the turbulent regime, by means of large eddy simulation (LES), a CFD-DEM model developed recently in our group for solid–liquid mixing. The resulting model is used to study the mixing of glass particles in a baffled stirred tank equipped with a down-pumping pitched blade turbine. Various characteristics of the liquid dynamics as well as the distribution and motion of the solids are investigated. The fraction of suspended solid particles predicted by the model is validated against experimental data obtained via the pressure gauge technique (PGT). Two new methods to calculate the fraction of suspended particles in a Euler–Lagrange simulation, the so-called Lagrangian suspended fraction analysis (LSFA) and the decorrelated fraction analysis (DFA) techniques are introduced. The results obtained with these two methods, as well as with many others taken from the literature, are compared to the Zwietering correlation and to the results obtained by the PGT. It is found that some techniques proposed in the literature, namely the local concentration, the power consumption and the transient solids concentration analysis techniques, cannot be applied adequately in this case. On the other hand, the LSFA, DFA and PGT techniques are observed to predict accurately the fraction of suspended solids when compared to experimental PGT data.

CFD-DEM modeling of cuttings transport in underbalanced drilling considering aerated mud effects and downhole conditions

This paper presents a developed CFD (Computational fluid dynamics)-DEM (Discrete elements method) model to study the cuttings transportation in aerated mud drilling process for inclined annuli at downhole conditions. The model is conducted to determine the effects of liquid flow rate, air injection rate, annulus inclination angle, elevated temperature and pressure on the cuttings transport efficiency. The motion of the fluid is computed using CFD based approach with gas–liquid interface capturing provided by the volume-of-fluid (VOF) method. The dynamics of cutting phase is studied by DEM using soft sphere approach in order to take into account the particle collision phenomenon. The continuum and dispersed phases are strongly coupled via the interaction forces such as drag force, lift force and pressure gradient force, which are considered in the developed CFD-DEM model. In addition, the real gas law governs the dynamics of injected air and the gas-liquid viscosity is assumed to be temperature dependant. The predicted results have been compared with the experimental data and fair agreement has been established. The results have indicated that in addition to the liquid flow rate, the cuttings transport efficiency is affected by gas-liquid ratio, ambient temperature and injection pressure.

Comparison of gasification performances between raw and torrefied biomasses in an air-blown fluidized-bed gasifier

A CFD-DEM model already developed by the authors has been extended to directly compare the gasification performances between raw and torrefied grassy biomasses in an air-blown fluidized-bed gasifier. The bed material is non-calcined olivine which acts as the solid heat carrier. Furthermore, effects of four key operating parameters (i.e., gasification temperature Tr, excess air ratio λ, steam addition, and biomass feed location) are also systematically examined. The results are analyzed both qualitatively and quantitatively by various indices: the fluidization behavior, bed pressure drop, product gas concentration profiles, total gas yield, and carbon conversion (CC). For both raw and torrefied biomasses, increasing Tr can enhance both the total gas yield and CC; rising λ decreases the H2 yield but increases the CO2 yield and CC; the steam addition has a positive influence on the total gas yield and CC and it can also be used to adjust the H2/CO ratio in the product gas; both the total gas yield and CC decrease with raising the height of the biomass feed location. For all cases, the torrefied biomass obtains a lower H2 yield and CC but a higher CO yield than its raw counterpart under the same operating conditions, suggesting that torrefied biomass requires a longer conversion process compared with raw fuel. Moreover, the gasification behavior of torrefied biomass is more dependent on the fuel feed location than raw fuel and such knowledge is important for the optimal design of fluidized-bed gasifier for torrefied feedstock.

Simulation of the transport and placement of multi-sized proppant in hydraulic fractures using a coupled CFD-DEM approach

The productivity of fractured wells is mainly governed by propped fractures, so it is of significant importance to find out where the injected proppants go during hydraulic fracturing treatments, as this is essential to scheduling proppant injection in fracturing design. Using a coupled CFD-DEM model, the transport and placement of multi-sized proppants in fractures in vertical and horizontal wells were systematically investigated, and the effects of having multi-sized particles relative to uniformly-sized ones on the proppant placement were quantitatively characterized. When a proppant-laden fluid is injected into a fracture in vertical wells, a small proppant bank quickly forms. The injected large and small proppant particles are almost uniformly mixed, with just a small-proppant region at the back side of the bank. In comparison in horizontal wells, a proppant dune first forms near the wellbore in a fracture, and the large proppant particles are more likely to accumulate near the wellbore while the small particles are transported deeper into the fracture. The main transport mechanisms of proppant particles are settlement and fluidization, which cause a three-layer flow pattern (stationary proppant bed, fluidization layer and clean fluid layer) to form. In addition, vortex is also an important proppant transport mechanism, especially in a fracture in horizontal wells, where the vortex drags the injected proppant particles to different locations causing a dual-dune profile. The effect of fracture tip screen-out on the proppant placement was investigated. Screen-out can significantly change the flow field in a fracture and this will subsequently affect final proppant placement. Ultimately, the process of graded proppant injection was realistically modeled, which shows small proppants to be transported deeper into the fracture, while large proppants accumulate near the wellbore.

CFD-DEM simulations of a fluidized bed crystallizer

In the present study, important features of the two-phase flow in a fluidized bed crystallizer are examined by numerical computations and companion experiments. The simulations are carried out using a coupled CFD-DEM approach (CFD: Computational Fluid Dynamics; DEM: Discrete Element Method). After validating an open-source CFD-DEM software tool for this purpose, regions within the crystallizer with unfavorable hydrodynamic features and thus a negatively impacted process outcome have been identified. This was first accomplished by single-phase CFD simulations. Then, the validated CFD-DEM model delivers valuable information that is difficult or even impossible to measure experimentally with sufficient accuracy, such as the velocity and position of fluidized crystals within the crystallizer.Since the simulations are computationally challenging, a compromise between simulated process time and number of simulated particles must be found. Hence, the CFD-DEM simulations are not utilized to simulate the whole crystallization process, but to examine a short time-window in detail. Corresponding findings confirm proper fluidization of the crystals support the model reduction carried out in a parallel project.

Comparison of the implementation of three common types of coupled CFD-DEM model for simulating soil surface erosion

Soil erosion is a common process studied by soil science, environmental engineering, geotechnical engineering, coastal engineering, and many other fields. In the areas of hydraulic engineering, the geotechnics of soil erosion remains a high priority topic as the bridge scour is a common cause of bridge failures. Accurate predictions of scour depth and soil erosion rate remain challenging, due to the limitations of existing scaled experimental approach in fulfilling the hydrological and hydrodynamic similarity requirements. Computational model offers a promising alternative to further the microscale understanding of soil erosion which can help to develop engineering tools in practice. Computational model that couples Computational Fluid Dynamics (CFD) and Discrete Element Method (DEM) to simulate the behaviors of fluid-solid systems is promising to advance the current tools for soil erosion analyses. Different mathematical forms for laminar fluid flows exist for the coupled CFD-DEM model as documented in published literatures and implemented in commercial and open-source software; each of them is based on certain physics assumptions and corresponding mathematical treatments. There are, however, no direct comparison of the results of CFD-DEM models based on these seemingly different mathematical formulations, which would help researchers to select the proper simulation tool. This study implemented coupled CFD-DEM models based on three most common types of mathematical formats used in the previous modeling work. The results of different CFD-DEM models are firstly validated by comparing the results of simulating the free settling of a particle in fluid. A case study is then designed to compare the models in simulating the surface erosion of cohesionless soil inside a pipe flow using laminar flow equations. Comparison indicates that for a relatively sparse particle-fluid system, the difference of the three models is negligible. For a dense particle-fluid system, simulation with the three different mathematical formats can predict different results (as large as 10% in the fluid velocity and 20% in the particle drag force for the simulation case study analyzed). The results of this case study indicate that the three CFD-DEM models achieved comparable results for simulating soil erosion from an engineering perspective, however, the differences between these models, which originate from their underlying physics assumptions, must be kept in mind in selecting an appropriate simulation model as well as in comparing the results from different models.

A coupled CFD-DEM model for simulating the rheology of particulate suspensions

A coupled CFD-DEM model for simulating the rheology of particulate suspensions

CFD-DEM investigation of viscous solid–liquid mixing: Impact of particle properties and mixer characteristics

In chemical engineering, numerous processes require the suspension of particles in a laminar or transitional regime. For such operations, predicting the fraction of suspended particles as well as their distribution and homogeneity is a major concern. In this work, the unresolved CFD-DEM model introduced by our group for solid–liquid mixing is used to investigate the mixing dynamics of viscous suspensions. The techniques chosen to characterize the degree of suspension, the homogeneity and the distribution of the particles are presented. They are used to assess the efficiency of a pitched blade turbine with a clearance of C=T/4. The impact of solid properties on mixing dynamics is investigated by varying the Young's modulus, the coefficient of restitution and the sliding friction coefficient in the DEM model. Lastly, five alternative configurations of the mixing rig are investigated by varying the clearance of the impeller and introducing baffles.