Capturing coalescence and break-up processes in vertical gas–liquid flows: Assessment of population balance methods

Gas–liquid flows are commonly encountered in industrial flow systems. Numerical studies have been performed to assess the performances of different population balance approaches – direct quadrature method of moments (DQMOMs), average bubble number density (ABND) model and homogeneous MUlti-SIze-Group (MUSIG) model – in tracking the changes of gas void fraction and bubble size distribution under complex flow conditions and to validate the model predictions against experimental measurements from medium- and large-sized vertical pipes. Subject to different gas injection method and flow conditions, bubble size evolution exhibited a coalescence dominant trend in the medium-sized pipe; while bubble break-up was found to be dominant in large-sized pipe. The two experiments were therefore strategically selected for carrying out a thorough examination of existing population balance models in capturing the complicated behaviour of bubble coalescence and break-up. In general, predictions of all the different population balance approaches were in reasonable agreement with experimental data. More importantly, encouraging results have been obtained in adequately capturing the dynamical changes of bubbles size due to bubble interactions and transition from wall peak to core peak gas void fraction profiles. As a compromise between numerical accuracy and computational time, DQMOM has performed rather well in capturing the essential two-phase flow structures within the medium- and large-sized vertical pipes when compared to those of ABND and homogeneous MUSIG models. From a practical perspective, the ABND model may still be considered as a more viable approach for industrial applications of gas–liquid flow systems.     

Numerical modelling of free surface flows in metallurgical vessels

A numerical model for simulating the transient behaviour of multi-fluid problems defined in 2D rectangular and cylindrical geometries is presented. The model uses a piecewise linear volume tracking scheme, and maintains sharp interfaces and captures fine-scale flow phenomena such as fragmentation and coalescence. The numerical model was applied to four problems of pyrometallurgical relevance – entrainment of matte in the flow of slag during skimming operations, splash resulting from a drop impinging on a bath, bubble rise in a liquid bath, and top-submerged gas injection. The numerical predictions are in good agreement with the published experimental results. The simulation of top-submerged gas injection showed, in detail, the phenomena of bubble formation, bubble rise, and splash drop formation and recoalescence with the bath. Data useful for engineering purposes such as pressure traces and time-averaged flow fields were obtained, allowing assessment of splash behaviour for given gas injection conditions. The numerical model has been shown to be versatile in being able to adapt to a wide range of multi-phase flow problems.     

污染物在非饱和带内运移的流固耦合数学模型及其渐近解

污染物在非饱和带中运移过程是多组分多相渗流问题.在考虑气相的存在对水相影响的前提下,基于流固耦合力学理论,建立了污染物在非饱和带内运移的流固耦合数学模型.对该强非线性数学模型采用摄动法及积分变换法进行拟解析求解,得出了解析表达式.对非饱和带内的孔隙压力分布、孔隙水流速以及污染物的浓度在耦合与非耦合气相条件下的分布规律进行解析计算.对该渐近解与Faust模型的计算结果进行了对比分析,结果表明:该模型解与Faust解基本吻合,且气相作用以及介质的变形对溶质的输运过程产生较大的影响,从而验证了解析表达式的正确性和实用性.这为定量化预报预测污染物在非饱和带中迁移转化和实验室确定压力-饱和度-渗透率三者之间的关系提供了可靠的理论依据.     

Effects of CZSi furnace modification on density of grown-in defects

During large diameter Czochralski silicon growth, heat zone and argon flow influence the formation of defects in silicon crystal by changing the distribution of temperature. Different silicon crystals with various density of grown-in defects were grown by replacing the popular heater with the composite heater and changing the popular argon flow into a controlled flow. The experimental results have been explained well by the numeric simulation of argon flow.     

Modelling gas injection of a Peirce–Smith-converter

A commercial CFD-code PHOENICS was used to solve isothermal flow field of gas and liquid in a Peirce–Smith-converter. An Euler–Euler based algorithm was chosen for modelling fluid dynamics and evaluating controlling forces of a submerged gas injection. Predictions were made with a k–ε turbulence model in the body fitted coordinate system. The model has been verified with a 1/4 scale water model, and a parametric study with the mathematical model of submerged gas injection was made for the PS-process and the ladle injection processes. Limits of the modelling technique used were recognised, but calculated results indicate that the present model predicts the general flow field with reasonable accuracy. Predicted bubble distribution, pattern of the flow field and magnitude of flow velocities were used to evaluate scaling factors of physical models and general flow conditions of an industrial PS-converter.     

A three-dimensional steady state thermal fluid model of jumbo ingot casting during electron beam re-melting of Ti–6Al–4V

A 3-D coupled thermal-fluid model describing mass, momentum and energy transport within a Ti–6Al–4V rolling ingot cast in an (Electron Beam Cold Hearth Remelting) EBCHR process has been developed to describe steady state casting conditions. The model incorporates a number of the physical phenomena inherent to the industrial process, including a metal inlet in the center of one of the narrow faces, complex boundary conditions based on industrial practice, buoyancy driven flow within the liquid and flow attenuation using a Darcy momentum source term within the mushy zone. The model ignores turbulence in the liquid pool and Marangoni (surface tension) driven surface flows. The model has been validated against liquid pool depth and profile measurements made on an experimental casting seeded with insoluble dense markers and doped with dense alloy additions. Comparisons have also been made to video images taken of the top surface during casting. The results indicate that the model is able to quantitatively predict the steady state sump depth and profile and is able to qualitatively predict aspects of the top surface temperature distribution. The model has also been used to conduct a process heat balance and sensitivity analyses. The process heat balance conducted on the model domain indicates that at steady state the liquid metal inlet contributes 88% of the total power input, while the electron beam provides net 12% after accounting for radiation losses from the top surface; 62% of the heat is lost through the ingots sides and the balance is lost via bulk transport of sensible heat through the bottom of the domain. The results of the sensitivity analysis on pool depth indicate that casting rate has the largest effect followed by metal inlet superheat. The thermal, flow and pressure fields predicted by the steady state model serves as the initial conditions for a transient hot-top model, which is the subject of a forth-coming paper.     

CFD modeling of hydrodynamic characteristics of a gas–liquid two-phase stirred tank

A three-dimensional CFD model was developed in this work to simulate hydrodynamic characteristics of a gas–liquid two-phase stirred tank with two six-bladed turbines and four baffles, coupling of the Multiple Size Group model to determine bubble size distribution. Important hydrodynamic parameters of the multi-phase system such as volume-averaged overall and time-averaged local gas holdups and axial liquid velocities along time and transversal courses were simulated and analyzed in detail, under varied operating conditions (inlet air flow rate and impeller rotation speed). Model predictions of local transient gas holdup and liquid velocity distributions on vertical and horizontal sections of the tank were also carried out. The overall flow patterns were discussed in detail to assess the mixing. Bubble size distributions were further predicted to reveal the unique properties of gas phase. Experimental measurements of overall gas holdups and local axial liquid velocities were used to validate the developed model.     

Turbulent particle dispersion in an electrostatic precipitator

The behaviour of charged particles in turbulent gas flow in electrostatic precipitators (ESPs) is crucial information to optimise precipitator efficiency. This paper describes a strongly coupled calculation procedure for the rigorous computation of particle dynamics during ESP taking into account the statistical particle size distribution. The turbulent gas flow and the particle motion under electrostatic forces are calculated by using the commercial computational fluid dynamics (CFD) package FLUENT linked to a finite volume solver for the electric field and ion charge. Particle charge is determined from both local electrical conditions and the cell residence time which the particle has experienced through its path. Particle charge density and the particle velocity are averaged in a control volume to use Lagrangian information of the particle motion in calculating the gas and electric fields. The turbulent particulate transport and the effects of particulate space charge on the electrical current flow are investigated. The calculated results for poly-dispersed particles are compared with those for mono-dispersed particles, and significant differences are demonstrated.     

The oxidation process of silicon: Modelling and mathematical treatment

In this paper we develop a mathematical model describing the oxidation process of silicon. At first we introduce a model with a sharp reaction front between the silicon and the oxide layer. From this we turn over to a phase field model where the reaction front is replaced by an extended reaction zone. The silicon dioxide and the silicon are regarded as components of a reacting mixture, and the oxygen is assumed to be dissolved in it. We formulate a local existence result of that second model in three space dimensions assuming some simplifications concerning the boundary conditions and solve the complete free boundary value problem numerically. For this, it has been implemented into the process simulator DIOS which allows two-dimensional computations.     

Heat and mass transfer in liquid desiccant air-conditioning process at low flow conditions

This paper investigates the transient heat and mass transfer in liquid desiccant air-conditioning process at low flow conditions. Using local volumetric average approach, one-dimensional non-equilibrium heat and mass transfer models are developed to describe the humid air and liquid desiccant interaction at counter flow configuration. Using triethylene glycol solution as desiccant, some experimental studies are completed. Experimental results are used to justify the numerical models. Numerical results are then obtained to demonstrate process characteristics. The models include a transient desiccant flow model for initial liquid desiccant building-up process, empirical wetted specific surface ratio for mass transfer, and heat and mass transfer coefficients. The objective of this research is to develop a process analytical tool for liquid desiccant air-conditioner design.     

多层气藏中气体流动问题的新模型及其应用

本文针对气井产量与井筒集是变数时,建立了多层气藏内真实气体渗流问题的新模型,求出了三种典型外边界条件下各储层压力分布精确解,作为特例,又得到了均质气藏内压力分布的精确解并给出了在气田开发中的应用.     

Gas-powder flow in blast furnace with different shapes of cohesive zone

With high PCI rate operations, a large quantity of unburned coal/char fines will flow together with the gas into the blast furnace. Under some operating conditions, the holdup of fines results in deterioration of furnace permeability and lower production efficiency. Therefore, it is important to understand the behaviour of powder (unburnt coal/char) inside the blast furnace when operating with different cohesive zone (CZ) shapes. This work is mainly concerned with the effect of cohesive zone shape on the powder flow and accumulation in a blast furnace. A model is presented which is capable of simulating a clear and stable accumulation region in the lower central region of the furnace. The results indicate that powder is likely to accumulate at the lower part of W-shaped CZs and the upper part of V- and inverse V-shaped CZs. For the same CZ shape, a thick cohesive layer can result in a large pressure drop while the resistance of narrow cohesive layers to gas-powder flow is found to be relatively small. Implications of the findings to blast furnace operation are also discussed.     

Multiphase transient flow model in wellbore annuli during gas kick in deepwater drilling based on oil-based mud

Transient-state gas and oil-based mud (OBM) two-phase flow in wellbore annuli will occur during gas kick. The phase behavior of influx gas and OBM will make the gas kick during OBM drilling more complicated. There are three possible cases in an annulus: only liquid flow in the entire annulus, gas and liquid two-phase flow in part of the annulus, and gas and liquid two-phase flow in the entire annulus. First, the phase behaviors of gas and OBM in wellbore annuli are studied based on the phase behavior of methane and diesel. A multiphase transient-flow model in annuli during gas kick based on OBM is then established based on gas–liquid two-phase flow theory and on flash theory in annuli. The influences of phase behavior in annuli and annular geometry are taken into account. The local flow parameters are predicted by the hydrodynamic models and the local thermodynamic parameters are predicted by the heat-transfer models in the corresponding flow pattern. The proposed model has a better performance, compared with two other models, against the published experimental data. Finally, the variation of pit gain, well-bottom hole pressure, and gas void fraction are obtained, leading to a better understanding of the occurrence and evolution mechanism of gas kick during deepwater drilling.     

Numerical prediction of recirculation flows with free convection encountered in gas-agitated reactors

Flows in a gas-agitated reactor have been predicted by a finite difference procedure. The free-convection phenomena in the gas-liquid mixtures have been accounted for by the calculation of a void fraction determined from the gas flow rate. Computations have been performed for two different situations: first, with the allowance of slip between gas and liquid phases, and second, without any slip. Reasonable agreement has been achieved between the measurements.     

Computational fluid dynamics modelling of gas jets impinging onto liquid pools

Gas jets impinging onto a gas–liquid interface of a liquid pool are studied using computational fluid dynamics modelling, which aims to obtain a better understanding of the behaviour of the gas jets used metallurgical engineering industry. The gas and liquid flows are modelled using the volume of fluid technique. The governing equations are formulated using the density and viscosity of the “gas–liquid mixture”, which are described in terms of the phase volume fraction. Reynolds averaging is applied to yield a set of Reynolds-averaged conservation equations for the mass and momentum, and the k–ε turbulence model. The deformation of the gas–liquid interface is modelled by the pressure jump across the interface via the Young–Laplace equation. The governing equations in the axisymmetric cylindrical coordinates are solved using the commercial CFD code, FLUENT. The computed results are compared with experimental and theoretical data reported in the literature. The CFD modelling allows the simultaneous evaluation of the gas flow field, the free liquid surface and the bulk liquid flow, and provides useful insight to the highly complex, and industrially significant flows in the jetting system.     

Numerical modelling of bubbly flows with and without heat and mass transfer

In this paper, modelling gas–liquid bubbly flows is achieved by the introduction of a population balance equation combined with the three-dimensional two-fluid model. For gas–liquid bubbly flows without heat and mass transfer, an average bubble number density transport equation has been incorporated in the commercial code CFX5.7 to better describe the temporal and spatial evolution of the geometrical structure of the gas bubbles. The coalescence and breakage effects of the gas bubbles are modelled according to the coalescence by the random collisions driven by turbulence and wake entrainment while for bubble breakage by the impact of turbulent eddies. Local radial distributions of the void fraction, interfacial area concentration, bubble Sauter mean diameter, and gas and liquid velocities, are compared against experimental data in a vertical pipe flow. Satisfactory agreements for the local distributions are achieved between the predictions and measurements. For gas–liquid bubbly flows with heat and mass transfer, boiling flows at subcooled conditions are considered. Based on the formulation of the MUSIG (multiple-size-group) boiling model and a model considering the forces acting on departing bubbles at the heated surface implemented in the computer code CFX4.4, comparison of model predictions against local measurements is made for the void fraction, bubble Sauter mean diameter, interfacial area concentration, and gas and liquid velocities covering a range of different mass and heat fluxes and inlet subcooling temperatures. Good agreement is achieved with the local radial void fraction, bubble Sauter mean diameter, interfacial area concentration and liquid velocity profiles against measurements. However, significant weakness of the model is evidenced in the prediction of the vapour velocity. Work is in progress through the consideration of additional momentum equations or developing an algebraic slip model to account for the effects of bubble separation.     

VLES turbulence model for an Eulerian–Lagrangian modeling concept for bubble plumes

Traditional Reynolds-averaged Navier–Stokes (RANS) approaches to turbulence modeling, such as the k-ϵ model, have some well-known shortcomings when modeling transient flow phenomena. To mitigate this, a filtered URANS model has been derived where turbulent structures larger than a given filter size (typically grid size) is captured by the flow equations and smaller structures are modeled according to a modified k-ϵ model. This modeling approach is also known as a VLES model (Very Large Eddy Scale model), and provides more details of the transient turbulence than the k-ϵ model at little extra computational cost.In this study a two-phase extension to the VLES model is described. A modeling concept for bubble plumes has been developed in which the bubbles are tracked as particles and the flow of liquid is solved by the Navier–Stokes equations in a traditional mesh based approach. The flow of bubbles and liquid is coupled in an Eulerian–Lagrangian model. Turbulent dispersion of the bubbles is treated by a random walk model. The random walk model depends on an estimation of the eddy life time. The eddy life time for the VLES model differs from a k-ϵ model, and its mathematical expression is derived.The model is applied to ocean plumes emanating from discharge of gas at the ocean floor. Validation with experiments and comparison with k-ϵ model are shown.     

A numerical model for predicting bubble formation in a 3D fluidized bed

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.     

Predicting gas–liquid flow in a mechanically stirred tank

Computational fluid dynamics (CFD) provides a method for investigating the highly complex fluid flow in mechanically stirred tanks. Although there are quite a number of papers in the literature describing CFD methods for modelling stirred tanks, most only consider single-phase flow. However, multiphase mixtures occur very frequently in the process industries, and these are more complex situations for which modelling is not as well developed. This paper reports on progress in developing CFD simulations of gas–liquid mixing in a baffled stirred tank. The model is three-dimensional and the impeller region is explicitly included using a Multiple Frames of Reference method to account for the relative movement between impeller and baffles. Fluid flow is calculated with a turbulent two-fluid model using a finite-volume method. Several alternative treatments of the multiphase equations are possible, including various expressions for drag and dispersion forces, and a number of these have been tested. Variation in bubble size due to coalescence and break-up is also modelled. The CFD simulation method has been used to model a gas-sparged tank equipped with a Rushton turbine, and simulation results are compared with experimental data. Results to date show the correct pattern of gas distribution and the correct trends in local bubble size in the tank. Further work is needed to improve the quantitative agreement with experimental data.     

On the distribution of throughput of transfer lines

Transfer lines simply characterise the interrelationship of manufacturing stages with their buffers and they are used to model the key features of such manufacturing environments with simplifying assumptions. There is a vast literature on these systems, however, little has been done on the transient analysis of the transfer lines by making use of the higher order moments of their performance measures due to the difficulty in determining the evolution of the stochastic processes under consideration. This paper examines the transient behaviour of relatively short transfer lines and derives the distribution of the performance measures of interest. An experiment is designed in order to compare the results of this study with the state-space representations and the simulation. Furthermore, extensions are briefly discussed and directions for future research are suggested.