CFD-DEM Approach Research Articles

Experimental and numerical investigation into iron ore reduction in packed beds

The objective of this contribution is to investigate into indirect iron reduction in a packed bed on both an experimental and a numerical level. For this purpose experiments of a packed bed of iron ore particles in a laboratory-scale reactor were carried out. Inflow conditions in terms of temperature and reducing gas composition were subject to change and the integral behaviour of the reactor was qualified by measuring the outflow conditions. In particular, the composition of the off-gas was analysed to determine the overall reduction degree of the packed bed. The numerical technique is based on a coupled DEM-CFD approach, in which the iron ore is treated as discrete particles and the flow of reducing gas is described by classical CFD. Each particle is characterised by its thermodynamic state, that is determined by solving one-dimensional and transient differential conservation equations for mass and energy. In conjunction with a reaction mechanism for iron reduction through carbon monoxide, the spatially and temporarily dependent reduction degree for each individual particle is resolved. Integrating over all particles yields the integral behaviour of the reactor. These results were compared to measurements and very good agreement was obtained.

Prediction of equilibrium mixing state in binary particle spouted beds: Effects of solids density and diameter differences, gas velocity, and bed aspect ratio

The state of solids mixing can be considered as a key factor in accomplishment of high efficiency and a uniform product in gas–solid fluidized-beds operating with several different particle sizes. In the present work, coupled DEM–CFD approach was used to carefully investigate the mixing process of binary particles in flat-bottom spouted beds. In this regard, simultaneous effects of particle specifications, operating conditions, and bed dimensions on the equilibrium mixing state of solid mixture were evaluated in detail. Moreover, taking into account the contribution of various parameters, the obtained simulation results were used to draw a correlation for predicting the equilibrium mixing index of similar spouted beds. In addition, the upper limit of particle specification difference and the appropriate bed dimensions to achieve a specific mixing index value was determined. Furthermore, the performance capability of spouted and conventional gas–solid fluidized beds regarding solids mixing was compared in detail.

Particle scale studies of heat transfer in a moving bed

Moving beds acting as bed reactors are quite common in industry. Understanding of the multiphase flow and thermal behaviour in moving beds is important for process design and optimisation. In this work, CFD–DEM approach is adopted to investigate the heat transfer behaviour in a moving bed which is relevant to the raceway region in an ironmaking blast furnace. The solid flow behaviour including flow pattern, velocity, and microscopic properties are investigated. A good agreement of gas temperature distribution between simulation and experiments is achieved. Then the heat transfer behaviour is analysed quantitatively through heat flux and contribution of each heat transfer mode. It reveals that under the conditions considered in this work, particle-fluid convection is dominant, and the contribution of radiation is very small but increases when the bed temperature is high. Solid flow rate and gas flow rate have significant but different influences on the momentum and thermal behaviour in moving beds. The study presented in this work provides a solid base for the further investigation of thermal behaviour in moving beds with the consideration of chemical reactions, and more complicated operational conditions.

A study of submarine steep slope failures triggered by thermal dissociation of methane hydrates using a coupled CFD-DEM approach

This paper presents an extended numerical study on submarine landsliding induced by instantaneous thermal dissociation of a methane hydrate layer, by considering different dissociation scenarios in terms of position and volume of dissociated gas hydrates. A novel scheme of coupling computational fluid dynamics and distinct element method (CFD–DEM) is implemented. The behavior of fluid is simulated by CFD incorporating an empirical equation of state for slightly compressible liquid, whereas that of sediment skeleton with methane hydrates by DEM through a two-dimensional thermo-hydro-mechanical bond contact model. The scenario of whole hydrate dissociation identifies four typical stages of submarine landsliding: the initiation of landslide after hydrate dissociation, the onset of the submarine landslide, the sliding process, and the final deposition and stabilization, which is characterized by some microscopic variables (e.g. particle displacement, fluid velocity, excess hydrostatic pressure, particle-scale energy input and dissipation) so to attain a complete analysis of the landslide process. Reasonable representation of the slope instability phenomenon as triggered by thermal dissociation of gas hydrates is obtained. Besides, slope failure scenarios with different positions and volumes of the occurring MH dissociation are presented to better understand their influence on the evolution of the instability. Four landslide types are generated by the different dissociation scenarios: fall, slump, flow and a combination of slump and flow, which are further analyzed by outputting some microscopic information (e.g. force-chain, averaged pure rotation rate and bond distributions).

CFD–DEM approach to investigate the effect of drill pipe rotation on cuttings transport behavior

Abstract Increasing cuttings bed height is a serious concern during extended-reach well drilling. In order to predict and prevent cuttings bed height increase, it is essential to study how the critical parameters influence the cuttings transport, especially the drill pipe rotation effects on the cuttings transport process. In conventional models for cuttings transport, the dynamic behavior of particles due to drill pipe rotation is neglected or empirically simplified. This paper presents a coupled Computational Fluid Dynamics and Discrete Element Method (CFD–DEM) approach to simulate the cuttings transport considering the dynamic collision process. The fluid phase is treated as an Eulerian continuum and described by CFD method. The cutting phase is modeled by DEM using a soft sphere approach for the particle collision dynamic. The continuum and dispersed phases are strongly coupled via the interaction forces such as the drag force, lift force and pressure gradient force, which is taken into account in the developed CFD–DEM model. The model takes into account the collisions of cutting–cutting, cutting–drill pipe and cutting–wall. Simulations are carried out for a number of laboratory-scale configurations, showing good agreement with experiment data reported in literature. The numerical simulations show that the drill pipe rotation builds cuttings presented in non-symmetric distribution along the hole and significantly decreases the cutting volume concentration at the medium and high fluid inlet velocities and lower range of drill pipe speeds. There is no additional contribution of drill pipe rotation after reaching a critical speed at high fluid inlet velocities. At low fluid inlet velocities, increasing the drill pipe rotation from 0 to 80 rpm dose not significantly affect the cuttings concentration; although an increase in drill pipe rotation speed from 100 to 120 rpm provided a remarkable improvement. Furthermore, at low inlet fluid velocities and horizontal inclinations up to 40°, drill pipe rotation can significantly decrease the cuttings concentration. The effect of drill pipe rotation on decreasing the cuttings concentration is almost negligible at high inlet fluid velocities for any inclination.

Coupled DEM-CFD Model to Predict the Tumbling Mill Dynamics

The charge motion in a tumbling mill has been mostly described by empirical, mechanistic and computational models. The computational model presented in this work is a three phase approach for the tumbling mill that combines a particle description for the solids modelled by the Discrete Element Method, and the continuum description for the fluid by CFD approach. In the present work a phase coupled approach is developed using C++ subroutines to model the charge and slurry dynamics inside a tumbling mill by mapping the particles on the CFD mesh and resolving the particle volume and velocities on per cell basis. The coupled DEM-CFD approach is implemented and the effect of drag force on the slurry by the particle motion. The set of coupled simulation were run varying the slurry rheology and results were validated with equivalent PEPT experiment of lab scale mills and a very good agreement is found in some cases. The Beeststra drag correlation was used to calculate the drag force between the fluid and the particles. The free surface profile of the charge as well as the slurry is calculated as well as the axial center of mass profile of the mill.

Numerical study of rope formation and dispersion of non-spherical particles during pneumatic conveying in a pipe bend

In pneumatic conveying bends are used to interconnect vertical and horizontal pipe sections. During conveying bends are known to be related to several characteristic flow phenomena. Experimentally, their investigation turns out to be difficult which favors the use of numerical approaches instead. Detailed investigations become possible by relying on Euler–Lagrange methods. Here, particularly combined Discrete Element Method and Computational Fluid Dynamics (DEM–CFD) approaches allow the transient description of the particle/fluid interaction. So far, only spherical particles have been considered by the DEM–CFD during pneumatic conveying with very few exceptions, e.g. [1], as DEM–CFD frameworks capable of representing non-spherical particles are not yet widely established. There is an ongoing discussion on proposed frameworks regarding applicable particle/fluid force models and the way of the particle shape representation. Due to these unresolved questions conveying of non-spherical particles through pipe bends involving phenomena like rope formation and dispersion has not been considered with numerical approaches such as the DEM–CFD, yet. Therefore, as many bulk solids involve complex shaped particles, rope formation and dispersion are investigated for non-spherical particles by a DEM–CFD approach for the first time. Exemplarily, cubes, octahedrons, pyramids, plates and icosahedrons are addressed. A DEM–CFD framework is developed which allows the modeling of arbitrary shaped particles. To underline the validity of the approach, results are benchmarked against established correlations which are available for certain particle shapes. Obtained results indicate differences in the pressure drop, particle velocity distribution, rope dispersion and the particle/particle, particle/wall and particle/fluid forces which are strongly dependent on particle shape.

Effect of the Friction, Elastic, and Restitution Coefficients on the Fluid Dynamics Behavior of a Rotary Dryer Operating with Fertilizer

Some fluid-dynamics aspects of a rotary dryer, operating with granular single superphosphate, were investigated experimentally and by coupled discrete element model–computational fluid dynamics (DEM–CFD) simulations. The experimental data of flight holdup and dynamic angle of repose in the upper half flights, at different angular positions, were successfully used to identify the best set of parameters for the spring–dashpot model used in DEM simulations. The smallest deviations from the experimental data were obtained in the simulation whose parameters values were elastic constant k = 400 N/m, friction coefficient μf = 0.2, and restitution coefficient η = 0.2. The results have shown that the coupled DEM–CFD approach with the selected parameters has been suitable to predict the fluid dynamics behavior of the rotary dryer operating with granular single superphosphate.

Insights in active pulsing air separation technology for coarse coal slime by DEM-CFD approach

To research a novel technology for dry coarse coal slime beneficiation and extend its application, active pulsing air separation technology was investigated by DEM-CFD coupling simulation approach. The results show that the ash content of feed is reduced by 10%–15% and the organic efficiency is up to 91.78% by using the active pulsing air separation technology. The gas-solid flow in the active pulsing air classifier was simulated. Meanwhile, the characteristics of particle motion and the separation process of different particles were analyzed, and the mechanical structure of the classifier was also modified to achieve high separation efficiency. Therefore, a novel high-efficiency dry beneficiation technique was advanced for coarse coal slime.

Investigating agglomeration phenomena in an air-polyethylene fluidized bed using DEM–CFD approach

Soft-sphere discrete element method and Navier–Stokes equations were coupled with the equations of energy to simulate the agglomeration process at high temperature in an air-polyethylene fluidized bed. The inter-particle cohesive force was calculated based on solid bridging by the viscous flow. The agglomerates were tracked as a real object with irregular shapes in which their translational and rotational motions were calculated according to the multi-sphere method. Fluidization behavior of an agglomerating bed was successfully simulated in terms of increasing the size of agglomerates as well as decreasing the bed height. The influence of gas velocity on the size of agglomerates was investigated and it was shown that the formation of massive agglomerates is postponed as the gas velocity is increased. Microscopic mechanistic studies in terms of contact time, cohesive force and repulsive force, which are the key parameters in the formation of agglomerates, were also carried out.

Modeling of Interior Ballistic Gas-Solid Flow Using a Coupled Computational Fluid Dynamics-Discrete Element Method.

In conventional models for two-phase reactive flow of interior ballistic, the dynamic collision phenomenon of particles is neglected or empirically simplified. However, the particle collision between particles may play an important role in dilute two-phase flow because the distribution of particles is extremely nonuniform. The collision force may be one of the key factors to influence the particle movement. This paper presents the CFD-DEM approach for simulation of interior ballistic two-phase flow considering the dynamic collision process. The gas phase is treated as a Eulerian continuum and described by a computational fluid dynamic method (CFD). The solid phase is modeled by discrete element method (DEM) using a soft sphere approach for the particle collision dynamic. The model takes into account grain combustion, particle-particle collisions, particle-wall collisions, interphase drag and heat transfer between gas and solid phases. The continuous gas phase equations are discretized in finite volume form and solved by the AUSM+-up scheme with the higher order accurate reconstruction method. Translational and rotational motions of discrete particles are solved by explicit time integrations. The direct mapping contact detection algorithm is used. The multigrid method is applied in the void fraction calculation, the contact detection procedure, and CFD solving procedure. Several verification tests demonstrate the accuracy and reliability of this approach. The simulation of an experimental igniter device in open air shows good agreement between the model and experimental measurements. This paper has implications for improving the ability to capture the complex physics phenomena of two-phase flow during the interior ballistic cycle and to predict dynamic collision phenomena at the individual particle scale.

Acceleration of coupled granular flow and fluid flow simulations in pebble bed energy systems

Fast and accurate approaches to simulating the coupled particle flow and fluid flow are of importance to the analysis of large particle-fluid systems. This is especially needed when one tries to simulate pebble flow and coolant flow in Pebble Bed Reactor (PBR) energy systems on a routine basis. As one of the Generation IV designs, the PBR design is a promising nuclear energy system with high fuel performance and inherent safety. A typical PBR core can be modeled as a particle-fluid system with strong interactions among pebbles, coolants and reactor walls. In previous works, the coupled Discrete Element Method (DEM)-Computational Fluid Dynamics (CFD) approach has been investigated and applied to modeling PBR systems. However, the DEM-CFD approach is computationally expensive due to large amounts of pebbles in PBR systems. This greatly restricts the PBR analysis for the real time prediction and inclusion of more physics. In this work, based on the symmetry of the PBR geometry and the slow motion characteristics of the pebble flow, two acceleration strategies are proposed. First, a simplified 3D-DEM/2D-CFD approach is proposed to speed up the DEM-CFD simulation without loss of accuracy. Pebble flow is simulated by a full 3D DEM, while the coolant flow field is calculated with a 2D CFD simulation by averaging variables along the annular direction in the cylindrical and annular geometries. Second, based on the slow motion of pebble flow, the impact of the coupling frequency on the computation accuracy and efficiency is investigated. It shows that the coupling frequency can be decreased to an optimal frequency without affecting the method fidelity. By analyzing one of the PBR systems, the HTR-10 design, we show that the combined acceleration strategies can reduce the simulation time by 80% while retaining the same accuracy as the tightly coupled 3D DEM and 3D CFD simulations. Therefore, it is expected that the proposed acceleration methods can advance the coupled DEM-CFD approach to analyze the PBR, accounting for more physics such as thermal and neutronic effects, on a routine basis.

Combined experimental and numerical approach for wear prediction in feed pipes

Within this work, a combined experimental and numerical approach to fundamentally understand erosive wear in feed pipes was initiated. By experimental lab-scale testing, it was shown that erosion rates strongly depend on the material's properties and testing conditions. Steel wear was more pronounced at higher impact angle, whereas low impact angle was more critical for rubber. Lab-tests results distinguish from empirical erosion models because material dependent critical impact energies and fatigue phenomena cannot be considered there. A CFD–DEM approach was conducted for simulation of particulate flow in pipes. In addition, long term wear measurements were done to gain data of the wear progress. Although further validation and testing are necessary, very promising results on erosion prediction could be achieved.

Modeling of the Spray Zone for Particle Wetting in a Fluidized Bed

AbstractIn a spray agglomeration process the particle wetting influences the agglomerate growth and particle dynamics in the granulator. The mass of binder liquid that is deposited on single particles affects the amount of energy dissipation during particle contacts. For the agglomeration of colliding particles the whole impact energy has to be dissipated due to viscous and capillary adhesion forces in the liquid film and plastic deformation of the material. Therefore, a detailed knowledge of the particle wetting is necessary to model the agglomeration process. This contribution uses a coupled DEM‐CFD approach to describe the spray zone of a two‐fluid nozzle in a fluidized bed agglomerator. Droplets modeled as discrete elements showed the formation of a spray zone with a conical shape. Simulations of the spray zone and the wetting of single particles are in good agreement with experimental results.

Numerical simulation of the liquid-induced erosion in a weakly bonded sand assembly

The paper presents a numerical study on the erosion of sand particles where hydrodynamic forces induced by liquid flow are considered to be dominant in a weakly bonded sand assembly under conditions relevant to sand production. The numerical approach employed is the coupled discrete element method (DEM) and computational fluid dynamics (CFD). The sand assembly modelled is composed of sand particles which are weakly bonded. Bond breakage occurs when the resolved normal and shear stresses in the bond exceed the bond strength. The disaggregated sand particles are then transported by liquid and considered as produced sand. The simulated results show that the main features of sand erosion can be captured by the CFD-DEM approach. It is also shown that the increase of axial compaction enhances sand erosion, and the increase in radial confining pressure can cause continuous sanding. The effects of different variables on erosion rate are examined, showing that particle–fluid interaction force is indeed the main driving force for sand erosion. Cavity formation is also examined by simply assigning a non-uniform bond strength distribution to the sand assembly. Cavities tend to form and grow within weaker regions. The results show that the proposed model offers a promising method, which should be further developed, to elucidate mechanisms governing the erosion of sand particles at a particle scale.

Characterization of solids mixing patterns in bubbling fluidized beds

Behavior of the solid phase in fluidized beds was studied by a 2D CFD-DEM approach to obtain more information on the solid mixing and circulation. Hydrodynamic parameters, including solid diffusivity, and internal and gross circulations were considered in this study. To validate the simulation, time-position data obtained by the Radioactive Particle Tracking (RPT) technique were used. It was shown that the 2D model can satisfactorily predict the axial diffusivity, while the radial diffusivity calculated based on the model is an order of magnitude smaller than the experimental one in 3D. The influence of aspect ratio of the bed, type of distributor, and inlet gas velocity on solids mixing pattern were also studied. The solids flow pattern in the bed changed considerably by increasing the aspect ratio. Different solid circulations were captured by numerical model for the two types of distributors, namely porous and injection types. The results suggested that increasing the superficial gas velocity caused rigorous internal and gross circulations, which in return, improved solids mixing and decreased deviations from well mixed state.

Improved DEM-CFD model and validation: A conical-base spouted bed simulation study

An improved and efficient DEM-CFD approach is developed for spouted beds. A nonlinear Discrete Element Method (DEM), with a concept of spring, dash-pot and friction slider, is used for tracing the movement of each individual particle. The gas flow is described by a set of reorganized governing equations. Two phases are coupled through contributions due to effects of porosity, viscosity and drag. All equations are solved with the commercial package Fluent with an implementation of User Defined Functions (UDF). To validate the improved model, a two-dimensional conical-base spouted bed is chosen as a case study. An unstructured mesh system is adopted instead of regular grid system. The simulation also takes the Saffman force and Magnus effect into account. The calculation results show good agreement with the experimental observations which are taken from the literature.

Discrete Element Method-Computational Fluid Dynamic Simulation of the Materials Flow in an Iron-making Blast Furnace

The cohesive zone, where the ore fed into the blast furnace softens and melts, is critical to the blast furnace performance and stability due to its influence on the gas and solid flow. Here we describe a project for the development of a process model to predict the cohesive zone properties and results of an important part of the work; the solid and gas flow models. The process model will be developed to describe a realistic solid burden flow and the formation of the cohesive zone, its shape, location, structure and permeability. This will be achieved using various simulation and computing tools: a combination of the Discrete Element Method (DEM) and Computational Fluid Dynamics (CFD), the coupled DEM-CFD approach, together with models for the thermodynamics and reaction kinetics. The key benefits of the coupled approach lie in the coupling of the continuous phase and the discrete particles, and the possibility of introducing thermodynamics and reaction kinetics into the system in a more realistic manner. DEM and coupled DEM-CFD simulations in several geometries are presented for reduced scale blast furnace investigation on the influence of non-spherical particles and gas flow on the solid flow. A large influence of the geometry shape and boundary conditions on the solid flow was also found.

Homogeneous and bubbling fluidization regimes in DEM–CFD simulations: Hydrodynamic stability of gas and liquid fluidized beds

In recent years coupled DEM–CFD models have been successfully utilized to simulate fluidized particle systems in the bubbling regime. In this paper we report on DEM–CFD simulations of liquid-fluidization of glass beads and gas-fluidization of Geldart's Group A particles carried out to characterize hydrodynamically the stability of the fluidized state, in the absence of cohesive forces. Due to the importance of the fluid–particle interaction, the two-way coupling approach used is introduced within a rigorous framework. Simulations of 200 μ m glass particles fluidized by water are presented. Homogeneous (particulate) fluidization is obtained for a wide range of velocities, in quantitative agreement with the theory of stability of the fluidized state and with experiments. The possibility of fine particles ( 70 μ m porous alumina, Geldart's Group A) to be air-fluidized homogeneously, without the need to impose inter-particle (cohesive) forces, is demonstrated by means of the first principles analysis guaranteed by the DEM–CFD approach. The appearance of bubbles in the fluidized bed behaviour is shown to occur at velocities in quantitative agreement with the theory of fluidized bed stability. In this context, the transient behaviour of the two systems considered in response to changes of fluid superficial velocity is analysed, allowing to validate the capability of the numerical model to capture the propagation of voidage shocks along bed height.