Search

The chemical looping hydrogen production (CLHP) process is a new approach for hydrogen ( H 2 ) fuel production. The proposed process consists of three reactors which are air reactor (AR), fuel reactor (FR) and steam reactor (SR). In this study, the gas solid flow behaviour in a cold model of the proposed CLHP process was simulated using computational fluid dynamics (CFD) with kinetic theory of granular flow. The effect of drag coefficient models on the pressure profile of each reactor was investigated and the results from the modified Syamlal–O’Brien drag model agreed well with the experimental data. The model was further used to investigate effects of the operating parameters on the hydrodynamics profiles of each reactor. The solid flux increased with the increasing of inlet velocity of AR and FR but decreased with the increasing of the inlet velocity of SR. At the same time, the solid flux increased with the increasing of the solid inventory.

Size-induced particle segregation in rotating drums has been investigated using an Eulerian continuum approach (ECA). The mixing process of a binary mixture were respectively predicted in an unbaffled drum, a drum with a “+” shaped central baffle and a drum with a “-” shaped central baffle. Qualitative comparisons with DEM results showed that ECA works well for describing the mixing process of binary mixtures. A mixing index based on sampling information from the computational grids was developed and used for quantitative analysis. Compared with a base case which uses equal-sized particles, the results showed that ECA can separate the effects of mixing from the effects of segregation. However, the rate of mixing were over-estimated and the improvement effects of “+” shaped baffle and “-” shaped baffle were hard to distinguish quantitatively in the ECA. The drag model was modified for improving the ECA. The final segregation was found independent on the initial particle distribution using ECA.

Discrete element method (DEM) has been used to investigate the effects of particle contact parameters, sliding friction coefficient (SFC), rolling friction coefficient (RFC) and restitution coefficient (RC), in a rotating drum. Their effects were examined in terms of mean power draw, mean collision energy and collision frequency employing response surface method (RSM) in conjunction with analysis of variance (ANOVA) to perform significance analysis for the effects of single factors and factor combinations. Various rotational speed was also considered in the present work. The results revealed that sliding friction coefficient was a determining factor for mean power draw and mean collision energy, and restitution coefficient had the greatest impact on collision frequency. Additionally, sliding friction coefficient and restitution coefficient showed interaction effects on both mean collision energy and collision frequency over the range of rotational speed. Moreover, the sensitivity of particle contact parameters varies with the rotational speed of the drum.

In continuum-based models, the kinetic theory of granular flow (KTGF) provides a model for calculating solids stresses but has limitations in densely packed regions such as mills, kilns and rotating drums. The Eulerian-Eulerian multiphase model coupled with KTGF is evaluated in the present work. An additional frictional stress models were added, and their suitability evaluated. For the evaluation, a rotating drum at three rotational speeds (20 rpm, 42 rpm and 65 rpm) was analyzed. Compared with Positron Emission Particle Tracking (PEPT) measurement data from literature, Johnson and Jackson's model and Schaeffer's model for the frictional stress both showed a lower angle of repose regardless of the wall boundary condition used. Thus, a new frictional viscosity model based on granular pressure was proposed. By adjusting the specularity coefficient of wall boundary condition, results of the present model agreed well with the PEPT measurements in terms of angle of repose and spatial velocity fields. In addition, considering that the actual Johnson and Jackson model for boundary condition includes two parts, collisional and frictional part, a discussion was made about boundary condition. The results showed that the validation of the proposed frictional viscosity model with experimental results could be completed at different rotational speeds by considering only the collisional part of the boundary condition with different specularity coefficients or only the frictional part of the boundary condition with different angles of friction or both parts with the same specularity coefficient and angle of friction. Nevertheless, when the complete Johnson and Jackson model of boundary condition was applied, the same specularity coefficient and angle of friction are used at different rotational speeds which is more physically meaningful. Moreover, it is found that the frictional contribution has greater influence on dynamic angle of repose than collisional contribution of the boundary condition in the current rotating drum.

A qualitative and quantitative investigation on the formation of scale in mixing tanks was conducted. The fabrication of a purpose-built tank that could be disassembled into nine segments (including a base, four walls, and four baffles or four blanking pieces) allowed the measurement of scale thickness in critical regions such as the impeller zone. The scale grown on the walls of the tank was physically scanned using a coordinate measuring machine (CMM). The CMM readings were then used to plot 3-D graphs of scale thickness and distribution on the walls of the tank. Experiments conducted under unbaffled and baffled conditions showed that the bottom region of the reactor was almost free of scale while the top region near the liquid surface had noticeable scale formation under both configurations. The scale thickness and distribution were lower in the unbaffled tank compared to the baffled tank, which, in line with CFD simulations, can be attributed to the increased tangential liquid flow velocity near the tank wall under unbaffled condition.

Rotary kilns are used in several minerals processing operations, as well as related industries such as cement manufacture. Since they involve complex multiphase heat and mass transfer processes, optimization would ideally be guided by numerical modelling. A new generic hierarchical approach is proposed in this paper, sequentially combining 2D Discrete Element Method (DEM) simulations of slices of the bed and 3D two-phase computational fluid dynamics (CFD) models of the entire kiln, encompassing both bed and gaseous freeboard. DEM simulations have been used to derive mean solids velocities in a bed for various rotation speeds and particle sizes, and these were then used to calibrate and validate the solids rheology model used in the subsequent CFD model. A modified Coulombic friction for the bed in the CFD model was shown to give satisfactory agreement with the DEM bed results over a range of rotation speeds. Similarly, thermal mixing simulations carried out using the 2D DEM model for the granular bed were used to validate a CFD model of heat transfer after calibration of the small-scale diffusive term. The capability of the resultant CFD model was demonstrated by 3D simulation of the bed of a large rotary kiln.

At present, the theoretical understanding of the collisional interaction between particles and a single bubble under well-defined laminar flow conditions is relatively sound. However, there is a lack of experimental data and predictions on details of the interaction processes in stirred turbulent conditions, as exists in mineral flotation cells. This paper investigates the effect of turbulence on the efficiency of collision between particles and bubbles from a fundamental point of view. A computational fluid dynamics method is used to establish a bubble–particle collision efficiency model system for turbulent flow. The 3D model is used to systematically study the effect of fluid turbulence on bubble–particle collision efficiency. In order to isolate the phenomena caused by turbulent flow physics, the study is conducted in two parts: large-scale turbulence and small-scale turbulence, which are respectively larger and smaller than the bubble size. Large-scale turbulence is assumed to increase the speed of bubbles relative to the slurry, and in most cases, increase in the speed increases collision efficiency, though in some cases the gravitational effect can cause decrease in overall efficiency. Small-scale turbulence is treated through the effect of fluctuations on particle trajectories. Again, such turbulence increases collision efficiency when the turbulent particle Stokes number is less than or about unity, but has little effect for Stokes number much greater than unity. An equation for the increase in efficiency in the limit of small Stokes number is obtained from the simulations.

Air–water models have been widely applied as substitutes for CO2–cryolite systems in the study of the complex bubble dynamics and bubble-driven flow that occurs in the molten electrolyte phase in the aluminum electrolytic process, but the detailed difference between the two systems has not been studied. This paper makes a numerical comparison between the bubble dynamics for the two systems. Simulations of both single bubble and continuous bubbling were conducted using a three-dimensional computational fluid dynamics (3D CFD) modeling approach with a volume of fluid (VOF) method to capture the phase interfaces. In the single bubble simulations, it was found that bubbles sliding under an anode in a CO2–cryolite system have a smaller bubble thickness and a higher sliding velocity than those in the air–water system for bubbles of the same volume. Dimensionless analysis and numerical simulation show that contact angle is the dominant factor producing these differences; the effects of kinematic viscosity, surface tension, and density are very small. In the continuous bubbling simulations, the continuous stream of air bubbles detaches from the anode sidewall after a period of climbing, just as it does in the single bubble simulation, but bubbles have less tendency to migrate away from the wall. Quasi-stable state flow characteristics, i.e., time-averaged bath flow pattern, turbulence kinetic energy, turbulence dissipation rate, and gas volume fraction, show a remarkable agreement between the two systems in terms of distribution and magnitude. From the current numerical comparisons, it is believed that the air–water model is a close substitutive model for studying bubble-driven bath flow in aluminum smelting processes. However, because of the difference in bubble morphologies between the two systems, and also the reactive generation and growth of bubbles in the real system, there will likely be some differences in bubble coverage of the anode in the anode–cathode gap.

Numerical simulations using two-fluid model incorporating the kinetic theory of granular flow (KTGF) are conducted to study the behavior of bed particles in a two dimensional baffle type internally circulating fluidized bed (ICFB). Complete knowledge of the solid circulation flux in an ICFB is very important as the bed particles are continuously circulating between different chambers above and below the central baffle. The effects of the gas distributor types, superficial fluidizing velocity and chambers height elevation difference are evaluated through computer simulations and the results are quantified in terms of solid circulation flux. By increasing the gas velocity (Uf) in the RC chamber, solid circulation flux from the heat exchange chamber (HEC) to the reaction chamber (RC) through the slot under the baffle is increased regardless of the distributor design. From further investigations, the solid circulation flux and particles radial velocity through the slot in the tubular gas distributor ICFB are found to be considerably increased as compared to the other configurations in our case. The elevated RC distributor is found to have a little impact on the solid circulation flux as a function of Uf. By varying RC distributor elevation while keeping HEC constant, the circulation flux is also changing and the highest flux is observed at an elevation level of 90 mm.

This paper modifies an experimental system previously used by Verrelli et al. to investigate particle–bubble attachment. Three different flow configuration cases were studied to investigate the effects of flow environment on the particle–bubble interactions and to model a situation more characteristic of real industrial froth flotation processes. The three configurations are as follows: particles dropped onto a bubble held stationary at the tip of a capillary tube; particles in water flowing past a bubble held stationary; and particles dropped onto a bubble rising freely. This is the first time collisional interaction between a freely rising bubble and hydrophobized glass ballotini particles has been studied in detail. The present work has shown the importance of considering real flow situations when modelling bubble–particle interactions for flotation applications. According to the results, there are more particles swept off the bubble surface before reaching the 90° position from the vertical axis under conditions of liquid flow (either because of externally generated flow or because of bubble rise), compared to the case where the bubble is stationary in a stationary liquid. This appears to suggest that attachment probability is reduced when liquid flows past the bubble. Extracting values of the individual force components provides greater insight into the mechanisms influencing the particle’s behavior as it interacts with the bubble. Comparison of particle trajectories that resulted in attachment with those that did not, provides further evidence in support of the proposition that contact time needs to be longer than induction time for attachment to occur.

A TFM–DEM hybrid model is introduced for modeling of the complex gas–solid flows in a pilot scale Coal Beneficiation Fluidized Bed (CBFB). The gas and the dense solid phases are modeled using an Eulerian-Eulerian or two fluid model (TFM), while the beneficiated coal particles are modeled as a dilute phase by the discrete element method (DEM). In this work, the influence of some key operation parameters on particle segregation behavior is studied, including fluidized air velocity, bed depth, and coal feed ratio and bed medium properties. Their effects are evaluated using a single coal sample of diameter 4.3 mm. Particles are divided into five different density fractions to represent the wide density range of raw coal samples. The simulation results demonstrate that by increasing the fluidizing air velocity from 1.2 u mf to 1.8 u mf of the dense medium solids, the segregation degree of beneficiated coal particles is significantly reduced, but the segregation time is only slightly decreased. Increasing the particle feed mass or decreasing the bed depth has a similar influence on CBFB operation. Both help to improve particle segregation, but a shallower bed is demonstrated to be more effective for coal beneficiation. A decrease in the medium density can reduce the bed cut density as well as the beneficiation limit for lighter samples, while a decrease in the medium size will increase the back-mixing effects, resulting in reduced beneficiation quality. Hydrodynamic forces acting on the beneficiated particles are also quantified from the simulation results. By analyzing the magnitude and direction of each force acting on discrete particles, the mechanisms influencing particle segregation under different operation conditions are explained at the particle scale.

This paper investigates the use of the multiphase particle in cell method (MP-PIC) in modelling the gas–solid flow in a laboratory-scale isothermal internally circulating fluidized bed (ICFB), and compares results with experimental observations previously published by Hadley et al. (“Experimental quantification of the solids flux in an internally circulating fluidised bed”, Fluidization XIII, 2010, pp. 885–892). Glass ballotini (140 μm mean diameter) were fluidized with room temperature air, and fluidisation velocities in both the central reaction chamber and two heat exchange chambers were varied to provide nine different chamber velocity ratio conditions for investigation. The model reliably predicted the overall bed dynamics of both gas and solid flow. Predicted solid recirculation rates were generally within one standard deviation of the mean experimental values, although under-prediction tended to occur at lower fluidisation rates. It was concluded that the MP-PIC method was a reliable tool for modelling bubbling fluidized bed behaviour of non-cohesive particles.

Particle segregation behavior in a coal beneficiation fluidized bed (CBFB) is numerically studied using a TFM–DEM hybrid model, in which the gas and the dense solid phases are modeled using a Eulerian–Eulerian or two fluid model (TFM), while the beneficiated coal particles are modeled as a dilute phase by the discrete element method (DEM). For validation purpose, the numerical model was setup using geometric and operating conditions similar to a laboratory experimental model with the bed thickness set to one particle diameter to save computational cost. For a fixed gas injection velocity, the influence of particle size and density of the beneficiated samples was studied. It was found that the particles would segregate along the bed height due to the density differences with the degree of segregation being strongly influenced by particle size. Obvious segregation occurs for the coarse samples (6.7 mm and 4.3 mm) and little segregation occurs for the particles smaller than 3 mm. The flow patterns and segregation kinetics were qualitatively comparable with those observed in physical experiments conducted under similar conditions. On this basis, the underlying mechanisms governing particle segregation have been explained in terms of the hydrodynamic forces acting on individual particles. It was demonstrated that the segregation of coarse particles was mainly controlled by the balance between gravity and the local pressure force, while fine particles were more strongly affected by the direct drag forces from the gas phase and the continuum solid phase, thus making them difficult to separate.

This paper reports on an assessment of the bubble-induced electrical resistance in the Hall-Heroult process for primary aluminium production through a combined use of physical and numerical modelling. Using a physical air–water model, the transient bubble dynamics beneath the bottom surface of an anode was captured using a digital camera. Bubble morphology information obtained from the experiment was used to set up a numerical model. Computational fluid dynamics (CFD) modelling was applied to predict the current flow and the corresponding voltage drop across the electrolytic cell with and without the presence of bubbles. The predicted bubble-induced voltage drop for a current density of 0.7 A cm−2 is about 0.11 V for a bubble coverage of 37 % and 0.29 V for a bubble coverage of 50 %. These values are within the range of experimental measurements reported for commercial cells. The predictions show that the presence of bubbles does not greatly affect global current distribution within the whole cell, but it does significantly affect the local current flow at the anode-bath interface. Locally high current flow occurs at the contact point of the anode bottom surface, bubble and liquid. In addition to the effect of bubble coverage, the bubble size and bubble thickness affect the voltage drop significantly.

A novel dry slag granulation process based on a spinning disc is being developed by CSIRO. This process utilises centrifugal force to break up molten slag into droplets, which are then quenched into solidified granules by a flow of cold air. In this process the sensible heat of slag is recovered as hot air. In the present work, a previously developed steady-state, two-dimensional and multiphase CFD model was applied to perform parametric numerical experiments to investigate the effects of a number of parameters on the liquid film thickness at the disc edge, which included liquid mass feeding (pouring) rate, disc spinning speed, disc radius, liquid viscosity, density and surface tension. The modelling results were compared with experimental data and were found to be in good agreement. To reduce the number of simulations needed, Box and Behnken's fractional factorial design of numerical experiment was adopted. Furthermore, in order for the modelling results to be applicable to atomisation of different liquids using spinning discs of different sizes, a dimensionless correlation was developed based on dimensional analysis of the numerical simulation data. The modelling results indicate that the liquid film thickness can be significantly influenced by the disc radius and spinning speed, the liquid mass feeding rate, viscosity and density, whereas the liquid surface tension has a negligible effect.

To improve understanding of the interaction between bubbles and particles during the flotation process, the interaction between the various factors that affect the collision efficiency has been analyzed in this paper. Four kinds of mineral particles (quartz, chalcopyrite, copper sulfide and galena) were investigated. Particle–bubble collisions between bubbles with size ranging from 0.6 mm to 2.0 mm and particles with diameter from 31 μm to 150 μm were examined. Representative analytical mathematical models (Generalised Sutherland Equation (GSE) model and Schulze model) are studied in depth, and errors in the equations for the GSE model reported in a review paper are corrected. Simultaneously, a computational fluid mechanics (CFD) technique was used to establish a fluid mechanics model which can directly reflect the relative motion of bubbles and particles during flotation. The results of the CFD model are compared with the existing mathematical models to analyze the advantages and disadvantages of the existing models as a description of collisions in the flotation process. The GSE model accounts for the centrifugal inertial effect (neglected in the Schulze model), but comparison with the CFD results suggests that it greatly over-estimates this effect. On the other hand, the Schulze model appears to overestimate the collision efficiency at the larger particle size range, giving values greater than unity in some cases. It would be expected that the CFD model would be more exact than either semi-theoretical model, given that it involves fewer assumptions. Therefore, it is recommended that the Schulze model (limited to a value less than unity) should be used as a sub-model for bubble–particle collision rate in macro-scale CFD models of flotation cells rather than the GSE model.

The supersonic separator is a novel compact tubular device for natural gas dehydration. The separation mechanism is not well understood for the complicated fluids with a delta wing located in the supersonic flows. We investigated the gas swirling separation characteristics in supersonic velocities using the Reynolds stress turbulence model. The results showed that the Laval nozzle designed with the cubic polynomial and Foelsch's analytical methods formed an extremely stable and uniform supersonic flow. The delta wing generated a strong swirling flow with the centrifugal acceleration of around 107 m s−2 to remove the condensed liquids from the mixture. However, the supersonic flow was quite sensitive to the delta wing, which led to the disturbance and non-uniformity of the gas dynamic parameters. This violent variation in the supersonic flow had a secondary action on the condensation, even resulting in the re-evaporation of the condensed liquids.

Aluminium metal production through electrolytic reduction of alumina in a cryolite bath is a complex, multi-physics, multi-scale process, including magneto-hydrodynamics (MHD), bubble flow, thermal convection, melting and solidification phenomena based on a set of chemical reactions. Through interactions of the different forces applied to the liquid bath combined with the different time and length scales, self-organised fluctuations occur. Moreover, the MHD behaviour causes a complex metal pad profile and a series of surface waves due to the meta-stable condition of the metal/cryolite interface. The large aspect ratio of an industrial cell, with a footprint of 20 by 4 m and at the same time having dimensions approaching just 30 mm of height for the reaction zone, prevents an integrated approach where all relevant physics are included in a single mathematical model of this large degree of freedom system. In order to overcome these challenges, different modelling approaches have been established in ANSYS® FLUENT®; Three models are used to predict details of specific physics: one to predict the electro-magnetic forces and hence the metal pad profile, a second that resolves details of the local bubble dynamics around a single anode and a third for the full cell bath flow. Results from these models are coupled to allow integration of the different phenomena into a full cell alumina distribution model. The current paper outlines each of the approaches and presents how the coupling between them can be realized in a complete framework, aiming to provide new insight into the process. This is a submitted manuscript of an article published by Elsevier Ltd in Applied Mathematical Modelling, 2 December 2016.

Supersonic quenching of magnesium vapour plays a pivotal role in the carbothermal reduction process for magnesium and the production of fine magnesium powders. Modelling of this process has previously been based on Classical Nucleation Theory in a one-dimensional flow expansion without considerations of background gas turbulence and the associated heat and mass transfers. This paper presents a single-velocity field, multi-component fluid model that overcomes the above shortcomings. The model has been validated using steam condensation data and applied to study supersonic quenching of magnesium vapour in a laboratory-scale Laval nozzle. Modelling results indicate a strong dependence of the vapour condensation characteristics on parameters such as vapour concentration and choice of carrier gas. The model is potentially a useful tool for designing and up-scaling processes that utilise supersonic quenching of metallic vapours.

This paper presents a numerical study of gas and solid flow in an internally circulating fluidized bed (ICFB). The gas and solid dynamics has been calculated using the commercial computational fluid dynamics (CFD) software package ANSYS/Fluent and an Eulerian–Eulerian model (EEM) with kinetic theory of granular flow used to calculate solid stresses. A two dimensional geometry was used to represent key parts of a laboratory ICFB. Simulations were conducted to assess the effect of changes to four designs or operating parameters: gas distributor plate angles, presence of a heat exchange tube bundle, superficial fluidizing velocities and initial solid packing heights. The mechanism governing the solid recirculation in an ICFB has been explained based on gas and solid dynamics obtained from the simulations and the effect of the four designs or operating parameters is quantified in terms of solid recirculation rate (SRR). Both of the investigated operating parameters, superficial fluidizing velocities and initial solid packing height, strongly affect solid circulation rate. For the two investigated design parameters, the presence of a tube bundle reduced the solid recirculation rate by 20%, while an inclination angle of 1.5° does not affect the solid recirculation rate significantly.

A major challenge in the use of brown coal in Victoria, Australia, for power generation is its high moisture content, which results in high greenhouse gas emissions. Predrying technologies for coal are one option being considered for use in brown-coal-fired power plants in Victoria to reduce greenhouse gas emissions. Using a validated computational fluid dynamics (CFD) model, this study investigates the combustion of predried brown coal in a 375 MW tangentially fired furnace that was designed for raw or non-predried brown coal. Different operating arrangements of the fuel gas for the predried coal with various moisture contents are proposed and assessed. When predried brown coal is burned with 55% or lower moisture content, the CFD results indicate that additional gas needs to be added to the fuel gas to maintain the original mass flow rate and reduce the heat flux in the furnace. Arrangements with additional air or recirculated flue gas added to the fuel gas are proposed and tested for predried coal with...

This paper presents a numerical study of a gas-particle flow in three inclined rectangular jets in crossflow. The predicted gas phase velocities and particle phase velocities are validated against previously reported experimental data. Two turbulence models, the standard k–ɛ model and Shear Stress Transfer (SST) model, are used to model the gas phase turbulence. This work shows that both models provide acceptable predictions of the gas flow and mixing generated by the three jets. Neither model could accurately reproduce the jet core and the flow near bottom wall. The particle phase in this flow comprises a large number of small particles. Thus particles follow the gas phase flow closely and any errors in the turbulence model and gas flow predictions are passed on to the particle phase simulation. This paper also includes a literature review on rectangular jets in crossflow and gas-particle laden jets in crossflow.

Fluidized bed technology has diverse industrial applications ranging from the gasification of coal in the power industry to chemical reactions for the plastic industry. Due to their complex chaotic non-linear behaviour understanding the hydrodynamic behaviour in fluidized beds is often limited to pressure drop measurements and a mass balance of the system. Computational fluid dynamics has the capability to model multiphase flows and can assist in understanding gas–solid fluidized beds by modeling their hydrodynamics. The multiphase Eulerian–Eulerian gas–solid model, extended and validated here improves on the kinetic theory of granular flow by including a closure term for the quasi-static stress associated with the long term particle contact at high solid concentrations. Similar quasi-static models have been widely applied to slow granular flow such as chute flow, flow down an incline plane and geophysical flow. However combining the kinetic theory of granular flow and the quasi-static stress model for the application of fluidized beds is limited. The objective of the present paper is to compare two quasi-static stress models to the experimental fluidized bed data of Bouillard et al. [4] . A quasi-static granular flow model (QSGF) initially developed by Gray and Stiles [18] is compared to the commonly used Srivastava and Sundaresan [37] . Both models show good agreement with the experimental bubble diameter and averaged porosity profiles. However only the QSGF model shows a distinct asymmetry in the bubble shape which was documented by Bouillard et al. [4] .

Gas vortices generated in the freeboard of a bubbling fluidised bed have become the centre of increasingly more research due to the advances in experimental technology. The behaviour of gas flow in the freeboard of a bubbling fluidised bed is of interest for applications such as the gasification of coal where reactions of gas mixtures, as well as gas–particle heat and mass transfer take place. Knowledge of the hydrodynamics of the gas within the freeboard can be hard to characterise, especially the detailed behaviour of gases escaping from bubbles that erupt at the bed surface. In the present study, experiments were conducted on a rectangular three-dimensional gas–solid fluidised bed. The experiments used a particle imaging velocimetry (PIV) measurement technique to visualise and measure the gas flow within the freeboard after a single bubble eruption. A computational study was carried out using Eulerian–Eulerian, kinetic theory of granular flow approach with a quasi-static flow model and with LES used to account for gas turbulence. Results from a three dimensional simulation of the experimental fluidised bed were compared with experimental velocity profiles of gas flow in the freeboard of the gas–solid fluidised bed after a bubble eruption. The CFD simulations showed a qualitative agreement with the formation of the gas vortices as the bubble erupted. Consistent with experimental findings the CFD simulations showed the generation of a pair of vortices. However, the simulations were unable to demonstrate downward flow at the centre of the freeboard due to particles in free fall after a bubble eruption event was observed in the experiments. Velocity profiles from the CFD data are in reasonably good agreement with the characteristic trends observed in the experiments, whereas the CFD model was able to predict the gas vortices phenomena and the velocity magnitudes were over-predicted.

A computational fluid dynamics (CFD) model of a 375 MW brown-coal-fired furnace in the Latrobe Valley, Australia, has been developed using ANSYS CFX 12.0. To improve the model predictions, a coal combustion model that takes into consideration carbon monoxide reactions has been utilized in ANSYS CFX 12.0. A level of confidence in the current CFD model has been established by carrying out a mesh independence test and validation against the furnace gas exit temperature (FGET), concentration of flue gas components, total boiler heat supply, and the wall incident heat fluxes measured in the power plant. The validated CFD model is then applied to investigate the effects of several operating conditions at full load, such as different out-of-service firing groups and different combustion air distributions on the coal flame. It is found that the selection of out-of-service firing groups has a considerable effect on coal combustion in terms of high-temperature zone shape and location and distribution of incident ra...

This paper describes the mathematical formulation and modelling issues of a computational fluid dynamics (CFD) model of a 375 MW utility furnace. This tangentially-fired furnace is fuelled by high moisture content brown coal from coal mines at Latrobe Valley in Victoria, Australia. The influences of different turbulence models, particle dispersion, and radiation models on the CFD prediction are investigated. Two turbulence models, standard k-e model and Shear-Stress Transport (SST) model, provide similar predictions that are in good agreement with the plant data. The effect of particle dispersion on the prediction is found to be insignificant for this high-volatile brown coal. The predicted wall incident radiation flux based on two radiation models, namely, discrete transfer (DT) model and P-1 model are compared against power plant measurements. The comparison reveals that the DT model provides good prediction of the radiation profiles, while the P-1 model considerably under-predicts the wall incident radiation flux.

The aerodynamic development of fully turbulent isothermal jets issuing from rectangular slot-burners was modelled by obtaining a solution to the Reynolds averaged Navier–Stokes equations. A finite-volume method was used with the standard k–e , RNG k–e and Reynolds stress turbulence models. The slot-burners were based on physical models, which were designed to be representative of typical burner geometries found in tangentially-fired coal boilers. Two cases were investigated, in which jets from three vertically stacked rectangular nozzles discharged at 90° and then 60° to the wall containing the burner. The nozzle angle had little effect on jet centreline velocity decay, with the 60° nozzle showing a marginally higher rate of decay. The jets from the 60° nozzles were found to deviate slightly from their geometric axis slightly due to internal pressure redistribution in the flow at the nozzles. The simulations were validated against the physical models and were found to reproduce the flow field of the jets accurately with the Reynolds stress model producing the best results.

The capability of six two-equation Reynolds-averaged Navier-Stokes (RANS) models for simulation of a non-swirl coal flame in a pilot-scale furnace has been investigated. These turbulence models—the standard k-ϵ model, re-normalization group (RNG) k-ϵ model, modified k-ϵ model, Wilcox k-ω model, Menter k-ω model (also called BSL model), and Shear-stress transport (SST) model—are assessed with the use of measured gas phase velocity, temperature, oxygen, and carbon dioxide volume fraction data from the literature. Predictions of the standard k-ϵ model, RNG k-ϵ model, BSL, and SST model are generally in good agreement with the experimental data. The Wilcox k-ω model generally overpredicts O2 volume fraction and underpredicts CO2 volume fraction. The modified k-ϵ model yields results that have large discrepancies from measurements.

To help improve hygiene in a copper smelter building, a correlation-based computer model of the gas extraction system was developed and used in conjunction with experimental observations to investigate the effects of process parameters on extraction performance. The model predicted that 28% of the total flow originated from in-leakage, and reduction of the in-leakage would cause a nearly equivalent amount of infiltration through the converter hoods, thus reducing fugitive emissions. It was also found that improved operational practices (closing and maintaining hood gages, isolating converters on standby) could double the draft within the converter hoods while decreasing the total flow rate by nearly 10%.

The carbothermic reduction of magnesia to produce magnesium offers the potential of a lower energy and higher productivity route for metal production compared to existing industrial routes. The reaction of magnesia and carbon produces a magnesium and carbon monoxide vapor. Slow cooling of that vapor will allow the reaction to quickly revert and the prevention of this reversion reaction is a major technical challenge. Two main approaches can be taken to prevent reversion and allow recovery of the metal product: rapid quenching of the vapor and dissolving the magnesium directly in a suitable metal solvent before reversion can occur. The commercial viability of either carbothermic route to magnesium is closely connected to the physical chemistry of each system.

The kinetic theory for particles of equal mass, density, and size is well established. However, kinetic theory for particles of different density and size is yet to be established. In this paper, the balance laws and constitutive relations for a granular mixture with particles of different size and densities are derived. The computed results indicate significant changes when compared with the theory that is derived by assuming particles of equal size and density. A comparison of the proposed theory with the theories given by Manger [Modeling and simulation of gas solid systems in curvilinear coordinates. PhD Thesis, Telemark Institute of Technology, Norway, 1996] and Huilin et al. [Physica, A, 284 (2000) 265] has also been made.

A phenomenological model for the creation and transport of anodic gas bubbles in Hall-H´eroult cells has been developed, following a multiscale approach. The essential features of the modelling framework are reviewed and discussed in the present paper. The model covers the generation of molecular gas species through Faradays law, subsequent bubble nucleation, and the evolution of macroscopic bubbles which are treated by a volume of fluid model. Recently, the modelling framework has been extended to include several complex phenomena such as surface tension and wetting, bath chemistry, and variable flow properties. The modelling framework has currently been applied to a laboratory scale electrolysis cell setup. The results demonstrate that the essential properties are well represented over a large range of experimental conditions by the proposed approach.

The paper presents the results of a computational model of erosion in a fluidised bed and a corresponding erosion experiment. The experiment has been simulated using the CFX [CFX-F3D, Version 4.1, Flow Solver User Manual, Computational Fluid Dynamics Services, AEA Industrial Technology, Harwell Laboratory, Oxfordshire, UK, 1995] code with computational models of hydrodynamics (hydrodynamic model A and kinetic theory model) and erosion (Finnie and kinetic theory). The experiment has been conducted at room temperature using a horizontal acrylic tube immersed in a rectangular fluidised bed for a total of 126 h of run. Erosion measurements were made every 14 h at eight equally spaced positions around the tube. The results show an induction period of 42 h. Most of the wear occurred around the bottom of the tube with the maximum at an angle of about 45° from the tube bottom. The kinetic theory model predictions are in good agreement with the experimental results.

Dust lift-off from conveyors forms a significant environmental and operational problem for operators in the mining, power generation and process industries. One means of reducing dust lift-off is to provide airflow deflectors or other aerodynamic modifications to the conveyor. A computational fluid dynamics (CFD) model has been developed to take into account the effect of wind direction, velocity and conveyor guarding on the dust loss from conveyors. The model is developed in the framework of CFX4. Experimental measurements of dust lift-off from the surface of a bed of ore in a wind tunnel at different wind velocities are used to characterise the dust. Based on the experimental data a model for predicting the mass and particle size distribution lifted from the bed surface at different air velocities is developed. The dust loss model is coupled to a Lagrangian particle-tracking model to predict particle trajectories. Validation of the model is undertaken by comparing CFD predictions against wind tunnel test work and shows good agreement. Results are presented for a typical conveyor design. The combination of experimental and CFD modelling is found to be a powerful tool for analysing dust loss from conveyors and can be extended to stockpiles and other situations where dust loss is a problem. The model can readily be extended to account for heat and moisture transfer in beds of porous materials.

Thrombus formation and hemolysis have been linked to the dynamics of blood flow in rotary blood pumps and ventricular assist devices. Hemolysis occurs as the blood passes through the pump housing, and thrombi develop in stagnation and low-velocity regions. The predicted velocities, pressure, and turbulence quantities from the numerical simulation are used to identify regions of high shear stress and internal recirculation. A nimerical technique is described that simulates the hydrodynamic characteristics of a rotary blood pump with a flow rate of 6 l/min at a rotational speed of 3000 RPM. A computational fluid dynamics (CFD) code, CFX 4, is used to solve the time-dependent incompressible Navier-Stokes equations using a transient finite volume method and three-dimensional structured grids. The simulation utilized the sliding mesh capabilities of this numerical code to model the rotating impeller and examine the effect of blade shape on the hydrodynamic performance of the blood pump in terms of pressure rise, flow rates, and energy losses. The first impeller model has six straight channels; the second impeller has six backward-curved channels. The results for two impeller configurations are presented and discussed. The curvedpump design resulted in higher pressure rise and maximum shear stresses than the straight-channel one. In general the paper demonstrates that CFD is an essential numerical tool for optimizing pump performance with the aim of reducing trauma to the blood cells.

In the steam explosion research field there is currently considerable effort being devoted to the modelling of premixing. Practically all models are based on the multiphase flow equations that treat the mixture as an interpenetrating continuum. The solution of these equations is non-trivial and a wide range of solution procedures are in use. Some numerical aspects of this problem are addressed here. In particular, the effect of the differencing scheme for the convective terms are examined and the use of hybrid differencing is shown to cause qualitatively wrong solutions in some situations. Calculations are performed for the Oxford tests, the BNL tests, a MAGICO test and to investigate various sensitivities of the solution. In addition, it is shown that the use of a staggered grid can result in a significant error that leads to poor predictions of ‘melt’ front motion. A correction is given that leads to excellent convergence to the analytic solution. Finally, the issues facing premixing model developers are discussed and the fact that model validation is hampered more by the complexity of the process than by numerical issues is highlighted.

Fluidized bed systems have the potential to be widely used in the power generation, mineral processing and chemical industries. One factor limiting their increased use is the lack of adequate design techniques for scaling such systems. A model has been developed for simulating gas–solid fluidized bed plant. The model uses a multiphase Eulerian–Eulerian technique to predict the transient behaviour of fluidized bed systems. The commercial CFD code CFX is used as the computational framework for solving the discretized equations. To overcome the problem of accurate geometrical representation experienced in previous models a body fitted grid system is employed. The model is used to predict isothermal flow in a three-dimensional bubbling fluidized bed. Predictions of the three-dimensional model show bubble formation with gas bubbles or voids preferentially moving along the centre of the bed. Predicted behaviour is qualitatively consistent with experimental observations.

Numerous experimental and modelling studies of the premixing stage of a steam explosion are underway world-wide. While there is encouraging agreement between model results and experimental data for medium-scale tests, the agreement for small-scale tests has been less encouraging. In this paper, we show that this discrepancy can be accounted for by the need to use higher order differencing schemes for the convection terms in the conservation equations. When such schemes are used, the features of the experiment are reproduced. We present results obtained using various total variation diminishing schemes, quantify the effect of added mass and lift forces, and investigate the effect of interfacial drag modelling.

Fluidized beds are widely used in many process industries including the food and pharmaceutical sectors. Despite being an intensive research area, there are no design rules or correlations that can be used to quantitatively predict the solid mixing in a specific system for a given set of operating conditions. This paper presents a numerical study of the gas and solid dynamics in a laboratory scale fluidized bed coating process used for food and pharmaceutical industries. An Eulerian–Eulerian model (EEM) with kinetic theory of granular flow is selected as the modeling technique, with the commercial computational fluid dynamics (CFD) software package ANSYS/Fluent being the numerical platform. The flow structure is investigated in terms of the spatial distribution of gas and solid flow. The solid mixing has been evaluated under different operating conditions. It was found that the solid mixing rate in the horizontal direction is similar to that in the vertical direction under the current design and operating conditions. It takes about 5 s to achieve good mixing.

The aerodynamics of fully turbulent jets supplied from rectangular slot-burners was modelled using the Reynolds Averaged Navier–Stokes (RANS) model. Three different turbulent models were considered, such as standard k-ε, RNG k-ε and Reynolds stress turbulence models. The recessed-type nozzle geometry was investigated to determine the effect of burner geometry on jet development. The slot-burner was based on physical models, which were designed to be representative of typical burner geometries found in tangentially-fired coal boilers. The study was validated against the physical models. The detailed flow field obtained from the simulations was used to explain the aerodynamic development of jets in such burners. It was found that the addition of a recess section to the nozzle geometry introduced significant changes to the flow due to complex pressure and mixing fields being set up inside the recess, which altered the jets once they exited into the open atmosphere.

Carbothermal production has been recognized as conceptually the simplest and cleanest route to magnesium metal, but has suffered from technical challenges of development and scale-up. Work by CSIRO has now successfully demonstrated the technology using supersonic quenching of magnesium vapor (the MagSonic™ Process). Key barriers to process development have been overcome: the experimental program has achieved sustained operation, no nozzle blockage, minimal reversion, and safe handling of pyrophoric powders. The laboratory equipment has been operated at industrially relevant magnesium vapor concentrations (>25% Mg) for multiple runs with no blockage. Novel computational fluid dynamics (CFD) modeling of the shock quenching and metal vapor condensation has informed nozzle design and is supported by experimental data. Reversion below 10% has been demonstrated, and magnesium successfully purified (>99.9%) from the collected powder. Safe operating procedures have been developed and demonstrated, minimizing the risk of powder explosion. The MagSonic™ Process is now ready to progress to significantly larger scale and continuous operation.