CFX Multiphase 14.5 L07 Interphase Mass Transfer

Interphase Mass Transfer
14. 5 Release
Multiphase Flow Modeling

in ANSYS CFX
© 2013 ANSYS, Inc. 11-1 Release 14.5
Overview
• Introduction
• Equilibrium Phase Change Model
• Non Equilibrium Phase Change Models
– Thermal Phase Change Model
• RPI Wall Boiling Model
– Cavitation Model (Rayleigh Plesset Model)
– The Droplet Condensation Model
• Appendix
– Solidification Modeling

Introduction
• Interphase mass transfer can proceed by bulk exchange of
material between phases:
• boiling
• condensation
• sublimation
• freezing
• melting
• cavitation
• flashing
• Interphase exchange of mass may be either thermally or
mechanically driven

Phase Change Models
Phase Change Phenomena
• Thermal: • Mechanical:
– Driven by temperature – Driven by pressure differences
differences
• Applications:
• Applications: – Cavitation
– Condensation – Flashing
– Boiling
– Pool Boiling
– Wall Boiling
– Melting

Phasic Continuity Equation
• Mass transfer is represented by sources in the Phasic
Continuity equations

(ρ α rα )    (ρ α rα U α )  Γ α  S α
t
• The term Ga represents mass source per unit volume into
phase α due to interphase mass transfer :
Np
Ga   Gab
b 1
where Gab represents mass flow rate per unit volume from
phase β to phase α
• Sa = user specified mass sources
Secondary Sources
• The mass source terms (Gab ) affecting the continuity equations are
referred to as primary mass sources.
• Mass transfer also induces additional sources and sinks between
other transport equations. These are referred to as secondary
sources, or secondary fluxes, due to mass transfer.
• The default form of the secondary source terms for a transported
variable f is:
S M α β  Γ αβ
 
φ β  Γ βα φα 

Γ αβ φ β : mass transfer from phase β into phase α carries the bulk
conserved quantity φβ into phase α

Γ βα φ α : mass transfer from phase α into phase β carries the bulk
conserved quantity φα into phase β
Equilibrium Phase Change
Assumes phase change occurs instantaneously

Equilibrium Phase Change Model
• The equilibrium phase change model assumes that the

mixture of the two phases is in local thermodynamic
equilibrium.
• This means that the two phases have the same
temperature and that the phase change occurs very
rapidly.
• The model is especially suitable for flows of condensing
vapors (for example, wet steam or refrigerants) with
small liquid mass fractions but it can also be used for
melting or solidification problems as well.

Phase Diagram
• Requires knowledge of the phase diagram

of the material being modeled.
– Superheat region
– Saturation region
• To determine the mass fraction of the vapor,

or quality (x), the flow solver uses the lever rule:
0 ℎ𝑚𝑖𝑥 ≤ ℎ𝑠𝑎𝑡, 𝑙
ℎ𝑚𝑖𝑥 − ℎ𝑠𝑎𝑡, 𝑙
𝑥= ℎ𝑠𝑎𝑡, 𝑙 ≤ ℎ𝑚𝑖𝑥 ≤ ℎ𝑠𝑎𝑡, 𝑣
ℎ𝑠𝑎𝑡, 𝑣 − ℎ𝑠𝑎𝑡, 𝑙
1 ℎ𝑚𝑖𝑥 ≥ ℎ𝑠𝑎𝑡, 𝑣
• Temperature is backed off from the Enthalpy/Temperature Curve.

• All mixture properties are derived from enthalpy and pressure (for which we are
solving)

• Requires the consistent material properties for the two pure substance
states involved in the phase change, e.g. consistent reference states for
the enthalpy and entropy.
• The difference between the two reference enthalpies must equal the
latent heat at the selected reference temperature and pressure.
– Liquid Reference Enthalpy = 0 at (Tref, Pref)
– Vapour Reference Enthalpy = L at (Tref, Pref)
• The saturation properties i.e. saturation curve :
– the vaporization curve for vapor-liquid phase change
– the fusion line for liquid-solid phase change
– the sublimation line for solid-vapor phase change
• The saturation properties are defined as a material property of a
Homogeneous Binary Mixture (HBM).

• A Homogeneous Binary Mixture is a mixture of two states (solid,

liquid or vapor) of the same pure substance. Eg :
– H20l = Liquid Water
– H20v = Water Vapour (Dry Steam)
– H20vl = HBM of water and water vapour (Wet Steam)
• Select the Homogeneous Binary Mixture as your domain fluid.
• On the Fluid Models form, you should select one of the mixture
components to be the Equilibrium Fraction and the other to be
the Equilibrium Constraint.
• There are three methods available to set up saturation properties :
– Using a pre-calculated Real Gas Property .rgp file.
– Using the Real Gas equation of state library.
– Using a general setup

Wet Steam Example
• Rotor Stator Turbine
– Angular Velocity 5100 rpm
– 18,980 nodes: 15,648
elements
– Equilibrium phase change
with RGP properties
– Equilibrium phase change
with
• Ideal Gas vapour,
• Constant property liquid,
• Antione equation saturation data

Wet Steam Example
• Equilibrium Model Results
– Steam RGP file

Wet Steam Example
• Equilibrium Model Results
– Ideal Gas Vapour, Constant Property Liquid, Antione Equation

Practical Guidelines
• For steady state calculations initialise with care

– Try to initialise as all gas or all liquid, if necessary ramp boundary
pressures and/or temperatures up/down with iteration to
required value to achieve this
– You may need to under relax pressure to avoid oscillations
• The closer inlet conditions are to saturation the more unstable
the solution is likely to be to the pressure
– A small time step may make this worse as mass continuity will
generate larger pressure gradients

Non-Equilibrium Phase Change
Mass transfer at Finite Rate and does not occur instantaneously

Non-Equilibrium Phase Change
• Thermal Phase Change Model

– Supports liquid superheating and vapour sub-cooling.
– Mass transfer driven by finite rate inter-phase heat transfer.
• Cavitation Model
– Occurs when liquid pressure drops below saturation pressure.
– Mass transfer driven by difference in pressure and saturation
pressure.
• The Droplet Condensation Model
– Occurs when near-saturation two-phase flow undergoes rapid
pressure reduction leading to nucleation and subsequent
droplet condensation.

Non-Equilibrium Phase Change:
Thermal Phase Change Model

Thermal Phase Change Model
• Addition or removal of heat
to a phase induces a phase
change as the temperature rises
Tv
above or falls below the saturation
temperature Ti
• Energy balance at phase interface
• Interfacial mass transfer rates Tl

Interfacial Energy Balance
• Steady-state energy balance:
Interface
. . Ti
 m lv H i v
 m lv H i l Liquid
Tv
 qi v
 qi l  0 mlv Hi mlv Hi
l v
qi l
qi v
• Interfacial heat transfer rates:
qi  h v  Tv  Ti 
 l
v
Tl
q i l  h l  Ti  Tl  Vapour

Interfacial Energy Balance
• Balance equation: Interface
Ti
qi  qi Tv
mlv  v l Liquid
Hi v
 Hi l mlv Hi mlv Hi
l v
• Interfacial enthalpies  saturation qi l

qi v
values
• Heat transfer rates from empirical
correlations
Tl  l
• Different on continuous and
dispersed side Vapour

Condensation Example
• Experiment at Lehrstuhl für Thermische Kraftanlagen, TU München
• Condensation onto free surface for steam/liquid water
Temperature Profile Measured

as a Function of Height
Adiabatic wall
Steam Thermocouples
128
Water
Adiabatic wall
790

• Vapor: • Liquid:
• Saturated vapor at constant pressure • ‘Large’ heat transfer
coefficient
T v
 T sat P • Vapour  laminar
• Liquid  turbulent, k-
H v
 const. • Turbulence damping at
free surface
qi v
0
• No energy equation in vapor phase

Grid 12160, y+=40
Liquid volume fraction

• Vertical Temperature profile at thermocouple location
40
Vapour
30
Liquid
20
CFD, no condensation
Increase due to latent
heat from condensation CFD, condensation
10
Experiment
0
0 50 100 150 200
T, °C
Condensation Example: Stationary Droplet
• Initially subcooled liquid droplet in contact with saturated vapor, heats up as
condensation takes place until saturation temperature is reached
• Dimensionless temperature of droplet, θ, predicted as a function of dimensionless time, as
characterized by the Fourier number, Fo
• Vapor: • Liquid:
• Saturated vapor at constant pressure • Heat transfer coefficient based on:
44 16
Nu  
T  T  P  9 ln 1  4 
v sat
H  const. T l  T0
v 
T  T0
qi 0 sat
t
v
• No energy equation in vapour phase Fo  2
d
Condensation Example: Stationary Droplet
1.2
1.0
0.8
Analytical
CFX-5
Theta
0.6
0.4
44 16
Nu  
0.2 9 ln 1  4 
T l  T0

0.0 T sat
 T0
0.00 0.05 0.10 0.15 0.20 0.25
t
 2
FoFo
d

Thermal Phase Change: Modelling Issues
• Inhomogeneous Eulerian Multiphase.

– Can also be run using homogeneous momentum
• E.g. If dispersed phase is very small bubbles
• Nucleation/Seeding
– The model is developed for interior boiling/condensation.
– The model requires both phases to be present to trigger mass
transfer (nucleation)

Thermal Phase Change: Seeding Options
• Seeding Options
– (1) Seed the inlet flow or initial guess with vapour volume
fraction.
• Only works close to or above saturation temperature.
• Significant liquid subcooling causes rapid condensation of
seeded vapour volume fraction.
– (2) Include mass sources to represent homogeneous phase

change based on departure from equilibrium
– (3) RPI Wall Nucleation Model

Thermal Phase Change: Seeding Options
– (4) Set a minimum vapour volume fraction.

• May cause convergence difficulties.
– (5) Set a minimum volume fraction to be used in interfacial
area density calculation
• Option (5) implies a small unphysical mass non-
conservation when the volume fraction of a mass
losing fluid goes to zero.

Thermal Phase Change Model +
RPI Wall Boiling Model

• For subcooled flows with
superheated walls, standard thermal
phase change models for bulk
boiling/condensation will
underpredict rates – bubbles form
at tiny nucleation cavities on walls
and grow
• The RPI Wall Boiling model ,

originally developed by Podowski
and co-workers , provides a
mechanistic model for wall driven
boiling in subcooled liquids under
bubbly flow conditions

Regime Changes for a Boiling Liquid
subcooled bubbly annular spray

flow flow slug flow flow flow
ONB OSB RPI Wall Boiling Model

wall temperature
T
Tsat mean fluid
temperature
x
subcooled nucleate boiling
boiling (saturated boiling)

• Sub-Grid Scale Model for Nucleate Boiling
– Similar to Turbulent Wall Functions
– Similar to model originally implemented in CFX-4
– Includes improvements to achieve grid independence (Egorov and Menter)
– Plus usability improvements:
– All wall heat transfer BC’s supported:
• Specified Heat Flux,
• Specified Temperature,
• Specified Heat Transfer Coefficient)
– Boiling at CHT boundaries supported
• Uses lagged wall temperature, hence not as robust as domain walls

• Determines Heat Flux Partition at Wall:
– Q = Qc + Qq + Qe
– Qc = Convective Heat Transfer
– Determined by Turbulent Wall Function
QQ
– Qq = Quenching Heat Transfer
– Departure of a bubble from heated surface 
cooling of surface by fresh water.
– Qe = Evaporative Heat Transfer QE
– Determined by physical sub-models on the sub-
grid scale.
QC

RPI Wall Boiling: Sub Models
• Several Sub Models:
• Nucleation site density n  nref ((Twall  Tsat ) / Tref )1.805

(Lemmart and Chawla) nref  7.9376 m2 , Tref  10 K
• Bubble Departure Diameter (Tolubinsky d dep  d ref exp  Tsat  Tliq  Tref 
and Kostanchuk)
d ref  0.6 mm, Tref  45 K
4 g
• Bubble Departure Frequency f 
(Cole) 3d dep liq
0.8
• Bubble Departure Waiting Time t wait 
f
• Wall area fraction influenced by bubbles (clipped) A2 f  d dep
2
n
• Wall area fraction influenced by single phase A1 f  1 A2 f

convection
Qc  A1 f hc Twall  Tliq 
• Convective Heat Transfer (Turbulent
Wall Function)
• Quenching Heat Transfer Qq  A2 f hq Twall  Tliq 
t waitkliq liqCPliq
hq  2 f


Qe  3
d dep fnhlg
• Evaporative Heat Transfer 6
• Mesh Independence
– Default bubble departure diameter depends on a ‘near wall liquid temperature’
– Experimental correlation uses pipe centre value
– Old implementations used value at centre of control cell adjacent to wall (mesh dependent)
– ANSYS CFX uses logarithmic wall function to correct this to an estimated value at fixed yplus
value = 250 (user adjustable), hence mesh independent
– Similar correction procedure is applied to quenching heat transfer coefficient

• Sub-Model Defaults
– Originally tuned for pressurised water
– May need retuning for other situations
– Possible to change default model constants
– Possible to over-ride with user’s own sub-model, using CEL.
– Both may be done on a per-domain or per-wall basis.
• Other models typically applied for variables outside the

RPI model itself
– Bubble diameter in the bulk (usually assumed as a function of
local liquid subcooling)
– Included non-drag forces

Wall Heat Partitioning: Iterative Solution
• Heat Flux Partition determines wall heat flux as a complex non-linear function of
wall superheat
Q  Qc Twall   Qq Twall   Qe Twall 
• Specified Wall Temperature: Twall  Tspec
– Use specified Twall to determine individual wall heat fluxes
• Specified Heat Flux: Qwall  Qspec

– Compute Twall from Qspec using non-linear equation solution algorithm
– Use computed Twall to determine individual wall heat fluxes
• Specified Heat Transfer Coefficient: Qwall  hT0  Twall 

– Same as for Specifed Heat Flux.
• CHT Wall
– Same as for Specifed Temperature, using lagged interfacial temperature

Wall Boiling Setup Procedure: Domain Level
• If wall boiling is required it must be
defined first at the domain level
• Optional Sub-Model of Thermal
Phase Change Mass Transfer
• Default sub-models are assumed
unless altered by the user

Wall Boiling Setup: Wall Boundaries
• Wall boiling must be switched on
explicitly on each wall where is expected
to occur.
– Defined under Fluid Pair objects under
boundary details
– Option not available if fluid dependent
heat transfer BCs are requested
• Wall boiling will not occur if you forget to
do this!

User-defined Sub-models
• Sub-model defaults may be
overridden on the domain or the
boundary level
– May change model constants
– May proved user-defined sub-models
via CEL
– Domain level modifications apply on
all walls where wall boiling is defined,
unless overridden locally
– Boundary level modifications only
apply locally to the specific wall.

Current Wall Boiling Model Limitations
• Sub-Cooled Nucleate Boiling Model
– Not applicable to Film Boiling or Critical Heat Flux (CHF) regimes.
• Turbulent Flow
– All turbulence models supported in principal. Laminar flow not
supported.
– Robustness problems observed for fine near-wall meshes using
Automatic wall functions.
• Other Restrictions
– Wall boiling + radiation not supported.
– Wall heat transfer BC’s must be defined per wall, not per fluid.

Wall Boiling Validation
• Validation Test Case
– Subcooled Boiling in Pipe with Heated Wall Bartolomej et al. (1967,
1982) (Conxita Lifante, 2008)
– large number of experimental testcase conditions with data
– steam-water pipe flow with wall boiling
– liquid sub-cooling defined to have steam inception always at the
same wall height
– Different configurations were studied in the paper.
Main parameters:
– Mass inflow rate
– Pressure
– Wall heat flux
– Pipe diameter

Bartolomej Test Case: Description
• 2D axial symmetry,
steady simulation
• 1 degree extrusion R
• Specified heat flux at the wall
• Symmetry b.c. at planes and
axis
X=2m
• Inlet b.c. with given inlet mass
flow
• Outlet b.c. with average static
pressure
X=X/100

Bartolomej Test Case: Models
• Steam-Water 2-phase flow:
– Water: continuous phase
– Water Steam: disperse bubbles
– Liquid temperature dependent bubble diameter
• Equation-of-state:
– IAPWS-IF97 water/water steam properties
• Interfacial Transfer
– Grace drag
– FAD Turbulent dispersion force
– Tomiyama lift force
– Wall lubrication force (Antal, Tomiyama, none)
– Two resistance heat transfer model, with Ranz Marshall on the
contintinuous phase side
– RPI wall boiling model with default sub-models
– User defined Interfacial area density accounting for higher volume
fraction of the steam phase
Bartolomej Test Case: Conditions
R = 7.7 mm
Variable Value
P 4.5MPa
q=0.57MW/m2
R 7.7 mm
Z= 2 m
Gin 900 kg/(s m2)
0.57MW/m2
q
Subcooling 58.2 K
Gin=900 kg/(s m2)

Numerical Grids
Grid Grid1 Grid2 Grid3
# Nodes
20x150 40x300 80x600
(uniform)
Max y+ 264 133 69
Δt [s] 10-2 10-3 5x10-4

Multiphase Models
Advection Scheme High Resolution
Drag Force Grace
Lift Force Tomiyama
Turbulent Dispersion Force Favre Averaged Drag Force
Wall Lubrication Force Antal
Enhanced Turbulent Production

Sato
Model
Wall Boiling Model RPI

Results: Grid1
Axial development of water temperature and steam volume fraction

Results: Grid2

Results: Grid3

Comparison to Data and Grid Independence

Total Gas Content
Grid Gas Content
Grid1 0.0789
Grid2 0.0779
Grid3 0.0822

Effect of Parameter & Model Variation
Wall heat flux influence: Wall lubrication force model influence:

Reactor Safety : Boiling in Rod Bundle
• 3×3 rod symmetry section

from a nuclear reactor fuel
assembly with guide vanes
• Periodic BC’s at all sides
• Wall heat flux of
qwall = 106 W/m2
• Reference Pressure
p = 15.7 MPa
• Water inlet temperature
TInlet = 607K
(= 12K water subcooling)

Reactor Safety : Boiling in Rod Bundle
• Vapor VF = 25% isosurfaces

Vapor Volume Fraction in Channels

Cavitation Model

Cavitation

Vaporous and Gaseous Cavitation
• Vaporous cavitation
• Occurs when the local pressure falls below the vapor pressure of a liquid.
• This causing very rapid boiling of the liquid at cavitation nuclei located within the
fluid.
• As the vapor bubbles move to regions where the vapor pressure is exceeded, they
collapse or condense.
• This condensation can be accompanied by local hydraulic shocks and the
formation of high velocity microjets.
• This type of cavitation is responsible for the mechanical damage that can occur on
ship's propellers.
• Gaseous cavitation
• Occurs when the local pressure falls below the saturation pressure of the dissolved
gas in a liquid. Dissolved gas can come out of solution as bubbles.
• Gaseous cavitation is a diffusion process and is much slower than vaporous
cavitation.
• Gaseous cavitation is responsible for the bends, or the formation of nitrogen
bubbles in the blood of scuba divers who ascend too quickly.

Will Cavitation Occur?
• The tendency for a flow to cavitate can be characterized by

the Cavitation Number which examines the difference between
the inlet absolute pressure and the vapor pressure relative to
the dynamic pressure head:
Pabs - P v
Ca =
U2/2
• As Ca decreases, the tendency for cavitation to occur increases

Cavitation: Thermodynamics
• 1  2: Pressure drop in the liquid
• 2  3: Vapour generation (cavitation)
P 1
• 3  2: Pressure increase because of Pv
volume increase followed by
condensation 2
• 3’: Intermediate state because of non- 3’

equilibrium effects
3

Vaporous Cavitation: CFD
• Important cavitation effects to model in CFD
– Modification of the pressure field
• Absolute pressure reduction due to Bernoulli velocity-pressure
effects should be bounded by the liquid vapor pressure - effect
on computed drag, lift, etc.
– Presence of vapor in the flow domain and the locations where it's
formed
• Can give guidance into possibility/liklihood of damage to
surfaces by cavitation
• Cavitation effects currently outside the scope of CFD
– Actual modelling of bubble collapse and the pressure pulse and
microjet formation that results
– Direct simulation of the damage to surfaces resulting from
cavitation bubble collapse

Interfacial Mass Transfer
Glv  mlv A lv
Mass transfer rate per unit area  Interfacial area/unit volume
• Interfacial area/unit volume  interfacial area density

• Obtain from
– Assumption on flow morphology (spheres, …)
– Model transport equation

Vapour Mass Generation Rate
• Spherical bubbles N
dm v dR
mlv   v R, dR/dt
dt dt
• Mechanical force balance:
– Rapid process, thermal
effects neglected
– viscous, surface tension
effects neglected • Linear bubble growth rate
(neglect 2nd order term)
 d 2 R 3  dR 2  dR 2 Pv  P
ρl  R 2      Pv  P 
 dt 2  dt   dt 3 l
Rayleigh-Plesset Equation
2 Pv  P
mlv  v
3 l
Interfacial Area Density
• Spherical bubbles present
at volume fraction aV: N
av Vcv 3av Vcv

N  R, dR/dt
Vb 4R 3
N 4R 2 3av
Alv  
Vcv R

Vaporization and Condensation Rates
3anuc 1  av  2 Pv  P
Glv  Fvap v if P  Pv
R 3 l
3av 2 P  Pv
Glv   Fcon v if P  Pv
R 3 l
• Availability of nucleation sites decrease as vapor volume fraction
increases (neighbor sites could be absorbed by a single bubble). A
modified interfacial area density is applied for vaporization.
• Empirical calibration coefficients: Fvap (50) and Fcon (0.01)
• Nucleation site density anuc (5  10-4) and Radius RB (10-6 m)

Cavitation Examples

Cavitation @ Hydrofoil
Ca = 0.34
• 1° angle of attack
• Cavitation at midchord
• Cavitation Numbers = 0.34 - 0.43
Cavitation @ Hydrofoil
Ca = 0.84
• 4° angle of attack
• Cavitation at leading edge
• Cavitation Numbers: 0.84 - 1.0

Cavitation Example: Pump Inducer
• Geometry and data: LEMFI-ENSAM, Paris
• Bakir et al. (2004)

Cavitation @ Inducer
Cavitation starts at Cavitation region Cavitation crosses
leading edge on grows over to pressure side,
suction side blockage = performance drop

Cavitation @ Inducer
12
12
N = 1450 rpm Experimental
N = 1450 rpm Experimental
qv/qn = 1.03 qv/qn = 1.09 CFX 5.7
CFX 5.7 10
10
8 8
Head (m)
Head (m)
6 6
4 4
2 2
0 0
0 2 4 6 8 10 12 0 2 4 6 8 10 12
NPSH (m) NPSH (m)

Cavitation @ Injection System
Rail pressure up to 1600 bar

12
11
Mass Flow [g/s]
10
9
8 dgap = 0.424 mm
7
6
5 din/out = 1.0 mm
0 20 40 60 80
Pressure Drop [bar]
Data Coarse grid Medium grid Fine grid

Cavitation erosion after 45 minutes

Data
3D-RANS
DES


Modeling Cavitation with ANSYS CFX
• Saturation pressure must be set if it's
not defined by a Homogeneous Binary
Mixture. If the simulation is isothermal,
then a constant value should be used.
• Default values for condensation and
vaporization constant, nucleation
site density and bubble diameter
can be adusted
• Absolute pressure (used for evaluating
compressible fluid densities) will be clipped
to a minimum value given by the vapor
pressure
• For post-processing purposes, relative
pressure is clipped to zero absolute pressure
(Solver Pressure shows the true relative
pressure used by the solver)
Cavitation Best Practice Guidelines
• Start with a converged non-cavitating solution before turning on

cavitation mass transfer
• For simulations with a large amount of cavitation, it may help
to perform several simulations starting from a simulation
with a low tendency to cavitate
• The cavitation rate can be underrelaxed or the vaporization and
cavitation constants can be temporarily reduced to increase
robustness

The Droplet Condensation Model
OR
Non-Equilibrium Steam (NES) Model

Non-Equilibrium Steam Condensation
• CFX is widely used in the
turbomachinery industry
• Steam turbine design is an important
application area for CFX
• Low pressure stages experience
complex two-phase flow behavior
when droplet condensation occurs
• Thermodynamic and aerodynamic
performance affected
• Large machines – small efficiency
improvements can lead to very large
economic benefits

Non-Equilibrium Steam Model
• Useful for situations where a dry (or near-saturation) two-phase flow
undergoes rapid pressure reduction leading to nucleation and
subsequent droplet condensation.
• Differs from the Thermal Phase Change model in that the droplet
diameter is calculated as part of the model rather than as a user input.
This leads to improved accuracy. In order to do this a transport
equation for droplet number (Nd) is solved :

(ρ d N d )    (ρ d u d N d )  ρ d rc J d
t
where Nd : number of droplets per unit volume
Jd : nucleation source term with units as number of droplets per
unit volume per unit time

Nucleation Source Term (Jd)
• The droplet phase can enter through the inlet or appear through
homogeneous (volumetric) or heterogeneous (surface) nucleation mechanism.
• The nucleation source term (Jd) has the following form :
 G * 
J d  A exp  
 kT 
 g 
where A = constant determined by the particular nucleation model, ΔG* = the Gibbs free energy change
at the critical radius (Rd*) conditions, k = Boltzmann's constant, Tg= the supercooled vapor temperature.
• The nucleation source term (Jd) is added to the transport equation of droplet
number (Nd) and phasic continuity equations :

(ρ c rc )    (ρ c rc u c )  Sd  m * rc J d 
t

(ρ d rd )    (ρ d rd u d )  Sd  m * rc J d
t
where Sd = interphase mass term, m* = nucleated droplet mass based on the critical radius (Rd*)

Small and Large Droplets
• NES model can be used for both small and large droplets.
• Small and large droplets use different models for heat transfer and phase
change.
• For large droplets, the heat transfer and phase change models described by
the Thermal Phase Change model should be used.
• For small droplets (less than 1 µm) the Small Droplet heat transfer model is
appropriate; it sets the droplet temperature (Td) to
Rd *
Td  Ts - Tsc
Rd
where Ts = saturation temperature, Tsc = supercooling level in gas phase, Rd = the droplet diameter
• For small droplets, the interphase heat and mass transfer models include
the influence of the Knudsen (Kn ) number on the Nusselt number.
• The dependence is required because droplet sizes vary significantly from free-
molecular (in a non-continuum regime) to continuum.
2
Nu 
1  3.18 Kn

Equilibrium vs Non-Equilibrium Condensation
• Differs from the Equilibrium Phase Change model in that it does not
assume the flow to instantaneously reach equilibrium conditions, and
therefore implicitly includes losses due to thermodynamic irreversibility.
Moisture delay
Equilibrium Solution Non-equilibrium solution

Setup of Non-Equilibrium Steam Condensation
• Compressible Gas
Phase
– Ideal gas.
– Redlich Kwong, or
IAPWS
(recommended)
• Homogeneous
Momentum
– Optional.
– Recommended for
small drops O(1mm)
• Inhomogeneous Heat
Transfer
– Compulsory

NES: Validation for Supersonic Nozzle
Supersonic Nozzle (Moore et. al., 1973):

NES: Validation for Supersonic Nozzle
Supersonic Nozzle (Moore et. al., 1973):

NES: Validation for Steam Turbine
• Model steam turbine
– Equilibrium model
part-load design over-load
– Non-equilibrium model (4 1 1 1
separate liquid phases)

0.8 0.8 0.8
– Comparison to data 0.6 0.6 0.6
span[]
span[]
span[]
0.4 0.4 0.4
0.2 0.2 0.2
E32 E32 E32

0 0 0
0.7 0.8 0.9 1 1.1 0.7 0.8 0.9 1 1.10.2 0.4 0.6 0.8 1 1.2
p/p_ref p/p_ref p/p_ref
1 1 1
0.8 0.8 0.8
0.6 0.6 0.6
span[]
span[]
span[]
0.4 0.4 0.4
0.2 0.2 0.2
E32 E32 E32

0 0 0
0.8 1 1.2 1.4 1.6 1.8 2 2.2 0.8 1 1.2 1.4 1.6 1.8 2 2.2 1.2 1.6 2 2.4 2.8 3.2 3.6 4
ptot/p_ref ptot/p_ref ptot/p_ref
Courtesy „Workshop Turbomachinery“, Gdansk Sept. 2006

A. G. Gerber, R. Sigg, N. Sürken, L. Völker, M. V. Casey
NES: Validation for Steam Turbine
design
off-design
Courtesy „Workshop Turbomachinery“, Gdansk Sept. 2006

A. G. Gerber, R. Sigg, N. Sürken, L. Völker, M. V. Casey
NES: Validation
Rotor Tip Section (Bakhtar et. al., 1995)

Appendix
(Modeling Solidification and Melting)

Modeling of Solidification and Melting
• While still classed as thermal phase change phenomena, solidification
and melting differ from condensation and boiling, since the processes
involve conversion of a mobile fluid to a stationary solid or vice versa
• Typical CFD implementations of solidification and melting use

momentum resistance terms to freeze the motion of the solid phase,
where they are tied to local values of the mass or volume fraction of
the solid phase
• The most widely used solidification model in CFX uses a

multicomponent approach rather than true multiphase and has been
successfully applied to continuous casting processes

Multicomponent Solidification Approach
• Single-phase model employing mass fractions for the solid and
liquid components
• Inclusion of transient and convective latent heat terms in the
enthalpy equation
• Additional advection term in the enthalpy equation representing
advection of component enthalpies relative to the mixture
enthalpy (in “mushy” regions) – custom solver needed
• Use of tables or algebraic expressions relating liquid fraction to
temperature, etc.
• Darcy-based treatment of flow in mushy and solid regions
• Damping terms in the turbulence equations in the mushy and
solid region
• Can be used for static solidification and continuous casting
applications

Mathematical Model - Energy
• Phasic energy equations in terms of volume fractions:
( S rS hS )
– Solid:    ( S rS U S hS )    (rS k S TS )  I
t
– Liquid: ( L rL hL )
   ( L rL U L hL )    (rL k L TL )  I
t
• Interphase heat transfer term I
• Phasic enthalpy definitions include the latent heat

• Phasic energy equations in terms of mass fractions:

(  YS hS ) 
– Solid:    (  YS U S hS )    ( YS k S TS )  I
t S
– Liquid: (  YL hL ) 
   (  YL U L hL )    ( YL k L TL )  I
t L
• where
1
  rS  S  rL  L 
YS /  S  YL /  L
rS  S rL  L
YS  ; YL 
 

• Assuming thermal equilibrium of phases (common temperature) add

phasic energy equations (interphase terms cancel):
 (  h)
   ( U h)    (k T ) 
t
  (  (YL (U L  U )( hL  h)  YS (U S  U )( hS  h)))
• h and U are the mixture enthalpy and velocity:
h  YS hS  YL hL U  YS U S  YL U L
• k is the mixture thermal conductivity

• The mixture enthalpy equation can be written more compactly as:
 (  h)
   ( U h)    (k T )    ( YS (U  U S )(hL  hS ))
t
• The final term represents differential phasic advection of energy
• The solid/liquid mass/volume fractions are determined algebraically as
functions of temperature, e.g.:
m
 T  TS 
rL    , rS  1  rL
 TL  TS 

Mathematical Model - Lever Rule
T
LEVER RULE
TP
Partition coefficient:
Gradient of liquidus:
TL Composition:
LIQUID
TM Solute conservation:
Solid fraction:
TS
MUSHY
SOLID
TE
SOLID
Y
YS Y0 YL YE

Mathematical Model - Momentum
• Two additional effects need to be included:
– resistance to flow in mushy region
– forcing of velocity to zero or casting speed in solid region
• Both effects are included through a Darcy-like momentum source:
mL
SM  (U  U S )
K
• The permeability, K, is given by the Kozeny-Carman equation:
rL3
K  K0
(1  rL ) 2

Mathematical Model - Turbulence
• Turbulent diffusion of heat must be prevented in solid regions

• Sink terms in the turbulence equations ensure this:
mL (1  rL )2 mL (1  rL )2
Sk   3
k , S   3

K0 rL K0 rL

Example: Hot-top Continuous Casting of
Aluminium
inlet – molten aluminium alloy
meniscus - adiabatic
water-cooled
mould wall
outlet

Continuous Casting of Aluminium Alloy

Continuous Casting of Aluminium Alloy

Thin-slab Continuous Casting of Steel
inlet – molten steel
20°C superheat meniscus - adiabatic
water-cooled
mould wall -
user-specified
heat transfer
coefficient
water-spray
cooling of strand
surface -
user-specified
heat transfer
coefficient outlet

Submerged Entry Nozzle (SEN)

detail, one half shown

Convergence behaviour

Streamlines coloured with time

Mold wall temperature - front

hot spots are due to superheated
jets from nozzle

Mold wall heat flux – front

greater heat fluxes correspond
to the higher temperatures
from the inlet

YS=0.1 YS=0.5 YS=0.9

YL=0.9 YL=0.5 YL=0.1
Liquid- and solid-fraction isosurfaces

coloured by distance from mid-plane –
front view
Liquid mass fraction shows
development of solidified shell
cooling and solidification due
to counter-current flow of
solid and liquid components is
evident

Conclusions
• Solidification model uses standard multi-component framework of CFX
• Temperature dependence supported in all properties (e.g. CpL, CpS,
hSL, YL vs. T, k, , etc.)

CFX Multiphase 14.5 L07 Interphase Mass Transfer

Uploaded by

Copyright:

Available Formats

CFX Multiphase 14.5 L07 Interphase Mass Transfer

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

CFX Multiphase 14.5 L07 Interphase Mass Transfer

Uploaded by

Copyright:

Available Formats

Interphase Mass Transfer

Multiphase Flow Modeling

© 2013 ANSYS, Inc. 11-2 Release 14.5

© 2013 ANSYS, Inc. 11-3 Release 14.5

Phase Change Phenomena

© 2013 ANSYS, Inc. 11-4 Release 14.5

© 2013 ANSYS, Inc. 11-7 Release 14.5

• The equilibrium phase change model assumes that the

© 2013 ANSYS, Inc. 11-8 Release 14.5

• Requires knowledge of the phase diagram

• To determine the mass fraction of the vapor,

• Temperature is backed off from the Enthalpy/Temperature Curve.

© 2013 ANSYS, Inc. 11-9 Release 14.5

© 2013 ANSYS, Inc. 11-10 Release 14.5

• A Homogeneous Binary Mixture is a mixture of two states (solid,

© 2013 ANSYS, Inc. 11-11 Release 14.5

© 2013 ANSYS, Inc. 11-12 Release 14.5

© 2013 ANSYS, Inc. 11-13 Release 14.5

© 2013 ANSYS, Inc. 11-14 Release 14.5

• For steady state calculations initialise with care

© 2013 ANSYS, Inc. 11-15 Release 14.5

© 2013 ANSYS, Inc. 11-16 Release 14.5

• Thermal Phase Change Model

© 2013 ANSYS, Inc. 11-17 Release 14.5

© 2013 ANSYS, Inc. 11-18 Release 14.5

© 2013 ANSYS, Inc. 11-19 Release 14.5

© 2013 ANSYS, Inc. 11-20 Release 14.5

• Interfacial enthalpies  saturation qi l

© 2013 ANSYS, Inc. 11-21 Release 14.5

Temperature Profile Measured

© 2013 ANSYS, Inc. 11-22 Release 14.5

© 2013 ANSYS, Inc. 11-23 Release 14.5

Grid 12160, y+=40

Liquid volume fraction

© 2013 ANSYS, Inc. 11-24 Release 14.5

© 2013 ANSYS, Inc. 11-27 Release 14.5

• Inhomogeneous Eulerian Multiphase.

© 2013 ANSYS, Inc. 11-28 Release 14.5

– (2) Include mass sources to represent homogeneous phase

– (3) RPI Wall Nucleation Model

© 2013 ANSYS, Inc. 11-29 Release 14.5

– (4) Set a minimum vapour volume fraction.

© 2013 ANSYS, Inc. 11-30 Release 14.5

© 2013 ANSYS, Inc. 11-31 Release 14.5

• The RPI Wall Boiling model ,

© 2013 ANSYS, Inc. 11-32 Release 14.5

subcooled bubbly annular spray

ONB OSB RPI Wall Boiling Model

© 2013 ANSYS, Inc. 11-33 Release 14.5

© 2013 ANSYS, Inc. 11-34 Release 14.5

© 2013 ANSYS, Inc. 11-35 Release 14.5

• Nucleation site density n  nref ((Twall  Tsat ) / Tref )1.805

• Wall area fraction influenced by single phase A1 f  1 A2 f

• Quenching Heat Transfer Qq  A2 f hq Twall  Tliq 

© 2013 ANSYS, Inc. 11-38 Release 14.5

• Other models typically applied for variables outside the

© 2013 ANSYS, Inc. 11-39 Release 14.5

• Specified Heat Flux: Qwall  Qspec

• Specified Heat Transfer Coefficient: Qwall  hT0  Twall 

© 2013 ANSYS, Inc. 11-40 Release 14.5