This action might not be possible to undo. Are you sure you want to continue?
Instructor: André Bakker
© André Bakker (2002-2005) © Fluent Inc. (2002)
• Multiphase flow is simultaneous flow of:
– Materials with different states or phases (i.e. gas, liquid or solid). – Materials with different chemical properties but in the same state or phase (i.e. liquid-liquid systems such as oil droplets in water).
• The primary and secondary phases:
– One of the phases is continuous (primary) while the other(s) (secondary) are dispersed within the continuous phase. – A diameter has to be assigned for each secondary phase to calculate its interaction (drag) with the primary phase. – A secondary phase with a particle size distribution is modeled by assigning a separate phase for each particle diameter.
• Dilute versus dense phase:
– Refers to the volume fraction of secondary phase(s).
• Volume fraction of a phase =
Volume of the phase in a cell/domain Volume of the cell/domain
• Laminar versus turbulent:
– Each phase can be laminar or turbulent. – Fluid flow (primary phase) may be turbulent with respect to the secondary phase but may be laminar with respect to the vessel.
Why model multiphase flow?
• Multiphase flow is important in many industrial processes:
– – – – Riser reactors. Bubble column reactors. Fluidized bed reactors. Scrubbers, dryers, etc.
Typical objectives of a modeling analysis:
– Maximize the contact between the different phases, typically different chemical compounds. – Flow dynamics.
• Annular flow: continuous liquid along walls. • Droplet flow: discrete fluid droplets in a continuous gas. • Particle-laden flow: discrete solid particles in a continuous fluid. • Slug flow: large bubbles in a continuous liquid.Multiphase flow regimes • Bubbly flow: discrete gaseous bubbles in a continuous liquid. • Stratified and free-surface flow: immiscible fluids separated by a clearly-defined interface. gas in core. slug flow bubbly flow droplet flow particle-laden flow annular flow free-surface flow 5 .
Flow regimes: vertical gas-liquid flow Evaporator Q(m 3 / s) Superficial Velocity : vsg (m / s ) = A( m 2 ) Q(m 3 / s) Q(VVM ) = 60 V (m 3 ) 6 .
• Dilute or dense. • Laminar or turbulent. slug flow. etc. Predicting the transition from one regime to another possible only if the flow regimes can be predicted by the same model.g.Multiphase flow regimes • • • • User must know a priori the characteristics of the flow. bubbly flow. annular flow. Flow regime. Only model one flow regime at a time. • Secondary phase diameter for drag considerations. This is not always the case. e. 7 .
• Two-fluids theory (multi-fluids). – Dispersed phase in a continuous phase. • Discrete element method. – Solve the trajectories of individual objects and their collisions. • Fully resolved and coupled. – Eulerian models. – Particles do not interact. – Track individual point particles. inside a continuous phase. • Algebraic slip model. – Solve one momentum equation for the mixture. 8 Increased complexity . – Solve as many momentum equations as there are phases. • Lagrangian.Modeling approach • Empirical correlations.
• Four-way coupling: – Includes all two-way coupling. – No influence of particulate phase on the gas phase. • Two-way coupling: – Fluid phase influences particulate phase via aerodynamic drag and turbulence transfer. – Particulate phase reduces mean momentum and turbulent kinetic energy in fluid phase. – Particle-particle collisions create particle pressure and viscous stresses. 9 .Coupling between phases • One-way coupling: – Fluid phase influences particulate phase via aerodynamic drag and turbulence transfer.
Modeling multiphase flows • • • • What is the goal of the simulation? Which effects are important? Controlled by which hydrodynamic effects? Controlled by which other transport Model Specific phenomena effects? Lagrangian Dispersed Phase • All these factors influence which model to Algebraic Slip choose for the analysis. Eulerian Eulerian Granular Volume of Fluid Flow Specific bubbly droplet particle-laden Process Specific slug Separation annular Filtration stratified/free surface Suspension rapid granular flow Evaporation Reaction ? 10 .
Mass transfer. Dispersion and breakup. Particle-particle collision. Heterogeneous reactions. Turbulence.Physical effects in dispersed systems • Hydrodynamics: – – – – – – – – – – – – Change in shape. Heat transfer. Coalescence. Inversion. 11 • Other transport phenomena: . Change in composition. Particle-wall collision. Diameter.
1 mm 18µ g f D τ p = 5. 2 dpρp τp = d p = 0.Hydrodynamics: particle relaxation time • When does the particle follow the flow? • Time when particle reaches about 60% of the initial difference velocity.5 10−2 s ρ p = 1000 kg / m 3 µ g = 1 10 −5 Pa s f D = Re p C D / 24 12 .
7 E-6 – 1.1-5 mm 0.4 s 5.7 s τp = 18µ g 13 .1-5 mm ρp 600-2000 1 2000 2000 2 d pρp µg 2E-5 Pa s 1E-3 Pa s 2E-5 Pa s 1E-3 Pa s τp 1.5E-2 – 140 s 1E-3 – 2.Hydrodynamics: particle relaxation time Typical relaxation times in process applications dp Coal combustion Bubble columns Sand/air Sand/water 1-500 µ 1-5 mm 0.4E-3 s 5.5E-5 – 1.
2 14 .Multiphase formulation • Two phases Fluid Solids • Three phases Fluid Solids .1 Solids .
001 . • All particle relaxation times. • Algebraic Slip Mixture Model (ASM). • Particle-wall interaction always taken into account. particle-particle usually not. • Eulerian-Eulerian model (EEM). • Particle relaxation times < 0.Model overview • Eulerian/Lagrangian dispersed phase model (DPM). particle-particle usually not. • Particle-wall interaction taken into account. • Eulerian-granular model (EGM). 07/05/11 . • All particle relaxation times. • Both particle-wall and particle-particle interaction are taken into account.0.01 s. • Neither particle-wall interaction nor particle-particle are taken into account. • All particle relaxation times.
16 . single set of turbulence transport equations solved. • For turbulent flows. • This approach works well for flow fields where both phases generally flow in the same direction.Algebraic slip model (ASM) • Solves one set of momentum equations for the mass averaged velocity and tracks volume fraction of each fluid throughout domain. • Assumes an empirically derived relation for the relative velocity of the phases.
j ∂ ∂ µeff ( + ) + ρm g j + F j + ∂xi ∂x j ∂xi ∂xi ∑α ρ u k k =1 n r r k k .i k . j u 17 .ASM equations • Solves one equation for continuity of the mixture: ∂ ρ ∂ ( ρui ) + =0 ∂t ∂xi • Solves for the transport of volume fraction of one phase: ∂ 2 ∂um .i um .i ∂um .iα2 α + =0 ∂t ∂xi • Solves one equation for the momentum of the mixture: ∂ ∂ ∂P ( ρum . j = − + ∂t ∂xi ∂x j ∂um . j ) + ρmum .
ASM equations (2) • Average density: ρ m = α 1 ρ 1 + α 2 ρ 2 • Mass weighted average velocity: um = ρ 1 α 1 u1 + ρ 2 α 2 u 2 ρ1α1 + ρ 2 α 2 • Velocity and density of each phase: u1 . ρ 2 r • Drift velocity: u1 = u1 − um • Effective viscosity: µ eff = µ 1α 1 + µ 2α 2 18 . u 2. ρ1 .
2 ( ρ m − ρ p )d p 18µ f f drag f drag 1 + 0.15 Re 0.Slip velocity and drag • Uses an empirical correlation to calculate the slip velocity between phases.0175 Re if Re >1000 19 .687 if Re ≤ 1000 = 0. ∆u = u = aτ rel p ∂um a = ( g − (um ⋅ ∇um + )) ∂t τp = • fdrag is the drag function.
– No sedimentation. • No interaction inside dispersed phase.0.Restrictions • Applicable to low particle relaxation times < 0. – No countercurrent flow.01s. 20 .001 p f drag = τ . • One velocity field can be used to describe both phases. 2 ( ρ m − ρ p )d p 18µ f • One continuous phase and one dispersed phase.
sup >> UL. Liquid 2 < L/D < 20 UG.sup up to 50 cm/s Liquid Pool Sparger UG.Example: bubble column design Gas A bubble column is a liquid pool sparged by a process stream.sup Liquid/Slurry Inlet Gas Gas Inlet 21 .
1980) 22 .Bubble columns: flow regimes Bubbly Flow Churn-Turbulent Flow (“Heterogeneous”) Flow Regime Map (Deckwer.
23 . flow rates and residence time. heterogeneous. – Volume. – Flow regime: homogeneous. – Mixing time. kla ~ usgaρ lbσ cµ ld µ geDf∆ ρ g are commonly used. slug flow. Correlations of the form α ~ usgaρ lbσ cµ ld are commonly used. and correlations for each are available. Correlations are available for a limited number of systems.Bubble column design issues • Design parameters: – Gas holdup. – Mass transfer coefficient kla. Correlations of the form – Axial dispersion occurs in both the liquid and gas phase. Directly related to rise velocity.
• Furthermore. experimental correlations may not accurately reflect changes in performance when flow regime transitions occur.cont’d • Accurate knowledge of the physical properties is important. • Although good correlations are available for commonly studied air-water systems. especially the effects of coalescence and mass transfer affecting chemicals. these are limited to the ranges studied. • Correlations may not be available for large scale systems or systems with vessel geometries other than cylinders without internals. 24 .Bubble column design issues .
• Coalescence and breakup models are not yet mature. 25 . bubble column reactors are modeled using a single effective bubble size. – The death and birth of each bubble bin is solved from the above models. – The gas phase is composed of n bubble bins and share the same velocity as the second phase. • Statistical approach. – Application of population balance with two-fluid models with initial focus on gas-liquid. • Population balance approach. Solve equation for number density.Bubble size • At present.
Start with α =0. Initialization of gas volume fraction is not recommended. but exact value of reference density is not crucial.Model setup • • The flow in bubble column reactors is typically modeled as being transient. First order differencing for all variables except volume fraction (second order) and in time. Generally set to that of the lighter phase.8 for turbulence. Typical underrelaxation factors 0. Use PISO. The time step used in the transient calculation needs to be small enough to capture the phenomena of interest and large enough to reduce CPU time. • • • • • • • • 26 . and 0. Typically use standard k-ε model for turbulence. Don’t forget to turn on gravity:-) Turn on implicit body force treatment (in operating conditions panel). A typical time step would be of the order of 1E-3 times the gas residence time. Set reference density.5 for pressure.5 for volume fraction.2 for momentum.1 for slip velocity (with ASM model). 0. 0. 0.
• This model solves only one momentum equation for the gasliquid mixture. Iso-contours of gas volume fraction 27 .ASM • 3D modeling of dynamic behavior of an air-water churn turbulent bubble-column using the ASM model.Bubble column example . • Constant bubble size is used.
28 .Bubble column example .ASM Unstable flow in a 3-D bubble column with rectangular cross section.
• Top hydrofoil impeller provides good blending performance in tall vessels.Example .gas-liquid mixing vessel • Combinations of multiple impeller types used. Eulerian Gas-Liquid Simulation Animation 29 . • Bottom radial flow turbine disperses the gas.
• Appropriate for modeling fluidized beds. • Granular volume fractions from 0 to ~60%. standpipes. • Added mass and lift force. hoppers. • Several choices for drag laws. • Frictional viscosity based formulation for the plastic regime stresses. Appropriate drag laws can be chosen for different processes.Eulerian-granular/fluid model features • Solves momentum equations for each phase and additional volume fraction equations. pneumatic lines. 30 . and particle-laden flows in which phases mix or separate. risers. • Several kinetic-theory based formulas for the granular stress in the viscous regime.
and particle-laden flows in which phases mix or separate. • Several choices for drag laws.Eulerian-granular/fluid model features • Solves momentum equations for each phase and additional volume fraction equations. pneumatic lines. 31 . hoppers. standpipes. risers. • Granular volume fractions from 0 to ~60%. • Appropriate for modeling fluidized beds. Appropriate drag laws can be chosen for different processes.
Granular flow regimes Elastic Regime Regime Stagnant Stress is strain dependent dependent Elasticity Plastic Regime Slow flow Strain rate independent Soil mechanics Viscous Rapid flow Strain rate Kinetic theory 32 .
– seeks its own level (“bed height”).Fluidized-bed systems • When a fluid flows upward through a bed of solids. – assumes the shape of the containing vessel. The suspended solids: – has many of the properties of a fluid. 33 . • Particle sizes vary between 1 µ m and 6 cm. • Bed height typically varies between 0. Particle sizes between 10 µ m and 150 µ m typically result in the best fluidization and the least formation of large bubbles.15 m/s to 6 m/s. Addition of finer size particles to a bed with coarse particles usually improves fluidization. • Superficial gas velocities (based on cross sectional area of empty bed) typically range from 0. Very small particles can agglomerate.3m and 15m. beyond a certain fluid velocity the solids become suspended.
Fluidized bed example 34 .
• Adsorption-desorption. • Gas mixing. • Classification (removal of fines from gas or fines from solids). • Size enlargement or reduction. • Solids mixing.g. 35 . • Coating. • Noncatalytic reactions (both homogeneous and heterogeneous). hydrocarbon cracking). – Physical contacting: • Heat transfer: to and from fluidized bed.Fluidized bed uses • Fluidized beds are generally used for gas-solid contacting. Typical uses include: – Chemical reactions: • Catalytic reactions (e. temperature control. • Heat treatment. between gases and solids. between points in bed. • Drying (solids or gases).
Typical fluidized bed systems .1 Gas Gas and entrained solids Dust Freeboard Solids Feed Disengaging Space (may also contain a cyclone separator) Separator Dust Fluidized Bed Gas in Solids Discharge Windbox or plenum chamber Gas distributor or constriction plate Bed depth 36 .
2 Gas + solids Riser section of a recirculating fluidized bed Solids Gas Bed with central jet Uniform Fluidization 37 .Typical fluidized bed systems .
Fluidization regimes Um f Um b U Uch U Solids Return Solids Return Solids Return Gas Fixed Bed Particulate Regime Bubbling Regime Slug Flow Regime Turbulent Regime Fast Fluidization Pneumatic Conveying Increasing Gas Velocity 38 .
slugs. solids feeder. • There is no single design methodology that works for all applications. and discharges. gas distributor). probes. 39 . disengagement space. L/D for staging.Fluidized bed design parameters • Main components are the fluidization vessel (bed portion. recirculating. instrumentation. The design methodologies to be used depend on the application. solids feed. flow control. • Location of instrumentation. turbulent. solids discharge. Flow regime changes can affect scale-up. solids retention time. gas supply. • Flow regimes: bubbling. • Temperature and pressure control. dust separator. • Heat transfer and flow around heat exchanger components. space required for internal components such as heat exchangers). • Typical design parameters are bed height (depends on gas contact time.
particle surface roughness. and the particle size distribution. particle cohesiveness. Hence we need to fine tune the drag term to match the experimental data for minimum fluidization or minimum bubbling velocity. 40 .Fluidized bed . – and the minimum bubbling velocity.input required for CFD • CFD can not be used to predict the: – minimum fluidization velocity. • All of these effects are lumped into the drag term. • These depend on the: – – – – particle shape.
frequency and population – short circuiting – effects of internals – – – – – – – effects of inlet and outlets hot spots reaction and conversion rates mixing of multiple particle size residence times of solids and gases backmixing and downflows (in risers) solids distribution/segregation 41 . CFD then can be used to predict: – – – – bed expansion gas flow pattern solid flow pattern bubbling size.when to use CFD • If the drag term is tuned to match the minimum fluidization velocity.Fluidized bed .
Fluidized bed animation VOF of gas Vs = 6 cm/s 42 .
Bubble size and shape .validation Gidaspow (1994) 43 .
validation Gidaspow (1994) 44 .Bubble size and shape .
and a steady state solution may not be numerically feasible unless a diffusive turbulence model (e. k-ε ) is used. to match the minimum fluidization velocity. 45 . • Broadly speaking. • Higher order discretization schemes give more realistic bubble shapes. • Use small time steps (0. as do finer grids.g.1s) to capture important flow features.001 to 0. • Always tune the drag formula for the specific applications. the Gidaspow model is good for fluidized beds and Syamlal’s model is good for less dense flows.Solution recommendations • Use unsteady models for dense gas-solid flows (fluidized beds as well as dense pneumatic transport lines/risers). These flows typically have many complex features.
– Eulerian model solves one momentum equation per phase. 46 . • In next lectures we will discuss particle tracking and free surface models.Conclusion • Today discussed the: – The algebraic slip model solves one momentum equation for the mixture.
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue reading from where you left off, or restart the preview.