Advanced Lagrangian Particle

Tracking
14. 5 Release

Multiphase Flow Modeling

in ANSYS CFX
© 2013 ANSYS, Inc. 8-1 Release 14.5
 Overview
• Other particle forces
• Transient particle tracking
• Particle breakup models
• Particle mass transfer
• Particle erosion
• Particle-wall interactions and wall films

© 2013 ANSYS, Inc. 8-2 Release 14.5

 Other Particle Forces

• Electrostatic • Van der Waals adhesion

– Electric charge of particles – Adhesion between particles and walls
– Precipitation, – Nano particles
powder painting • Near-wall lift and drag
– Electrostatic particle forces can be – Non-homogeneous flow
modeled as a beta feature with the
– Drag and lift increase at small wall
add-on MHD module
distances y < 5dP
• Magnetic
• A desired particle force that is not
– Metal separation implemented in CFX can likely be set using
• Thermophoretic a particle user fortran routine - see the CFX
– Temperature gradients in plasma documentation and the User Fortran
flows folder in the Examples directory in the
installation folder
– MHD generators

© 2013 ANSYS, Inc. 8-3 Release 14.5

 Steady State vs. Transient Particles
• Steady State Particles • Transient Particles
– Particles are always tracked from “birth” – Particles are tracked from start to
to “death” end of each time step
– Particles must not be restarted – Particles must be restarted
– All particles are injected at once – A time step step specific
number of particles is injected
– One-way coupled particles are solved – One-way coupled particles are
at the end of the simulation solved at the end of each time
step
– Fully coupled particles are solved
– Fully coupled particles are solved at a
at a user specified coefficient loop
user specified iteration interval interval within each time step
– Particle traveling time and particle
– Particle traveling time and particle time are different
time are the same – Post processing of time-integrated
particle information

© 2013 ANSYS, Inc. 8-4 Release 14.5

 Transient Particles: Particle Integration

• Integration of particles over actual time step:

t F  t F  t P  t F
• Piecewise linear forward Euler integration:
 P ,t  t  f ( P ,t ,  F ,t )
p

t P t P t P
Time
t F t F
t F  2t F t F  t F tF

© 2013 ANSYS, Inc. 8-5 Release 14.5

 Transient Particles: Particle Integration

• Time/space interpolation of the fluid variables to the actual particle time and
position:
 F ,t  f ( F ,t ,  F ,t  t ,  F ,t  2 t )Vertex
F F

• Time interpolation can be controlled with an expert parameter:

– pt time interpolation (Default: 1)
– 0: use newest time level tF
– 1: linear interpolation (using time levels tF and tF -tF)
– 2: quadratic interpolation (using time levels tF, tF -tF and tF -2tF)
 requires second order transient scheme

t P t P t P
Time
t F t F
t F  2t F t F  t F tF
© 2013 ANSYS, Inc. 8-6 Release 14.5
 Integration Controls for Transient Particles
• Solver controls applied for coefficient loops
within each time step
• Optional controls for fully coupled particles:
– First Iteration for Particle Calculation (1)
– Iteration Frequency (1)
– Particle Source Target (0.01)
• Based on the rate of change of sources
– Under Relaxation Factors (1.0)
• Velocity Under Relaxation Factor
• Energy Under Relaxation Factor
• Mass Under Relaxation Factor
• Velocity Under Relaxation Factor for First Particle
Integration
• Energy Under Relaxation Factor for First Particle
Integration
• Mass Under Relaxation Factor for First Particle Integration
• Velocity Under Relaxation Factor at Time Step Start
• Energy Under Relaxation Factor at Time Step Start
• Mass Under Relaxation Factor at Time Step Start
© 2013 ANSYS, Inc. 8-7 Release 14.5
 Transient Particle Injection
• Transient particles can be injected on
boundary patches and particle injection
regions
• Same CCL syntax as used for steady-state
particles, except …
• Number of Positions:
– Direct Specifications
 Number per Unit Time, [s^-1]
– Proportional to Mass Flow Rate
 Number per Unit Time and
Mass Flow Rate, [kg^-1]

© 2013 ANSYS, Inc. 8-8 Release 14.5

 Transient Particle Injection
• Particles are injected at start of time step or are distributed over
the time step

• Logical expert parameter:

– pt distribute trans particles (Default=t)

1. Time step 2. Time step

Particles are
distributed over
the time step Time

Particles are
injected at start
of the time step

© 2013 ANSYS, Inc. 8-9 Release 14.5

 Transient Injection Example
• Spray injection
– Given particle mass flow distribution:

Mass flow
Injected particle
mass flow
1

0 1 2 3
CEL:
Time
FUNCTION: MFPART
Argument Units = [s] PARTICLE INJECTION REGION: PIR1
Option = Interpolation FLUID: Aluminium
Result Units = [kg s^-1] INJECTION CONDITIONS:
INTERPOLATION DATA: INJECTION METHOD:
Data Pairs = 0,0.0, 0.5,0, 1,1, 1.5,1, 2,0, 4,0 Option = Cone
Extend Max = No …
Extend Min = No NUMBER OF POSITIONS:
Option = One Dimensional Number per Unit Time and Mass Flow Rate = NPART
END Option = Proportional to Mass Flow Rate
END END
EXPRESSIONS: END
DT = 0.05 [s] …
MFPARTMAX = 1 [kg s^-1] PARTICLE MASS FLOW RATE:
NPART = 100/DT/MFPARTMAX Mass Flow Rate = MFPART(t)
… END
END END
END END

© 2013 ANSYS, Inc. 8-10 Release 14.5

 Transient Injection Example

Mass flow
• Simple injection into a straight duct
1

0 1 2 3
Time

© 2013 ANSYS, Inc. 8-11 Release 14.5

 New Particle Fates for Transient Flows
• New particle fates for transient particles (INTEGER counters:)
– Entered domain +--------------------------------------------------------------------+
| Particle Fate Diagnostics |
– Left domain +--------------------------------------------------------------------+
| Particle type | Fate type Particles |
– Collected on walls +--------------------------------------------------------------------+
| Water | Entered domain : 100 |
– Integration error | | Continue from last time step : 1436 |
| | Collected on walls : 104 |
– Exceeded distance | | Waiting for next time step : 1432 |
+--------------------------------------------------------------------+
– Exceeded maximum number
of integration steps
– Reached minimum mass
– Exceeded time integration
– Absorbed by Porous Media
– Continue from last time step
– Waiting for next time step

© 2013 ANSYS, Inc. 8-12 Release 14.5

 Transient Boundary Vertex Arrays
• Instantaneous particle vertex arrays (for each time step):
– <Particle Material>.Wall Mass Flow Density
– <Particle Material>.Erosion Rate Density
– …
• New for transient particles:
 Time integrated particle vertex arrays:
– <Particle Material>.Time Integrated Wall Mass Flow Density
– <Particle Material>.Time Integrated Erosion Rate Density
– …

• Add particle boundary information to trn-files:

– Expert parameter: pt bnd arrays in trn files = t

© 2013 ANSYS, Inc. 8-13 Release 14.5

 Boundary Vertex Arrays Example

Instantaneous Time Integrated

Wall Mass Flow Density Wall Mass Flow Density

© 2013 ANSYS, Inc. 8-14 Release 14.5

 Particle Break-up Mechanisms
Primary Break-up Secondary Break-up

Dense Spray Dilute Spray

Injection Nozzle

Liquid
Liquid Core Dispersed Flow

• Primary Break-up
– In-nozzle effects (cavitation, turbulence induced disturbances)
– Instabilities on liquid-gas interface lead to primary break-up
• Secondary Break-up
– Droplets become unstable under the action of forces induced by their
motion relative to the continuous phase
© 2013 ANSYS, Inc. 8-15 Release 14.5
 Primary Break-up Models

• Task:
– Determine starting conditions of droplets that leave injection nozzle
• i.e.: initial radius, velocity components, spray angle
• mainly influenced by nozzle flow
• Various approaches of different complexity exist in literature
• Models available in CFX
– Blob method
– Enhanced blob-method
– LISA
– Turbulence induced atomization

© 2013 ANSYS, Inc. 8-16 Release 14.5

 Turbulence Induced Atomization (TIA)
• First introduced with the 12.0 release
– Based on model proposed by Huh [1] with modification of [2]
– [1] Kang Y. Huh et al. Diesel spray atomization model considering nozzle exit
turbulence conditions. Atomization and Sprays, vol.8, pp.453-469, 1998.
– [2] C.A. Chryssakis et al. Development and Validation of a Comprehensive CFD
Models of Diesel Spray Atomization Accounting for High Weber Numbers. SAE
paper 2006-01-1546, 2006.
• Main Features
– Turbulence effects inside the nozzle determine
– initial droplet size
– droplet velocity after primary breakup
– Allows to predicts initial spray angle
• Most appropriate in combination with Reitz & Diwakar or TAB secondary breakup
models which are discussed later

© 2013 ANSYS, Inc. 8-17 Release 14.5

 Turbulence Induced Atomization (TIA)
• The turbulence generated in the
nozzle hole produces initial
perturbations on the jet surface when Detached droplet
it exits the hole
LA

• Perturbations grow exponentially due

to aerodynamic forces and form new
droplets
Lw  2 LA

Turbulence
in the nozzle flow

Wave growth
due to relative velocity

© 2013 ANSYS, Inc. 8-18 Release 14.5

 Turbulence Induced Atomization: Setup
• Atomization model is selected from Domain 
Particle Injection Regions Tab

• Select “Cone with Primary Breakup” option and

select the primary breakup model

• Main Input parameters

– Injection Pressure Difference across Injector
– Density Probe Normal Distance
– Injector Nozzle Length over Diameter Ratio
– Model Constants
– Nozzle Discharge Coefficient, CD
(Model is very sensitive to this parameter!)
– Discharge Coefficient (constant value)
Vp
CD   f (m P , p)
– Automatic 2p  P

© 2013 ANSYS, Inc. 8-19 Release 14.5

 Turbulence Induced Atomization: Validation
Bosch case 1

Hiroyasu&Kadota, case 1
0.1
0.1
Penetration Depth [m]

We = 170 We = 1800
0.08 0.08

Penetration Depth [m]

0.06
0.06

0.04
Experiment
Blob + Reitz&Diwakar 0.04
Huh + Reitz&Diwakar
0.02 Blob + TAB Experiment
Huh + TAB Huh
0.02 Blob + Reitz&Diwakar
0 Huh + Reitz&Diwakar
0 0.0005 0.001 0.0015 0.002
Time [s]
0
Bosch case 2 0.0005 0.001 0.0015 0.002 0.0025
Time [s]

Experiment
0.1 Blob + Reitz&Diwakar
Huh + Reitz&Diwakar
Blob + Reitz&Diwakar
Penetration Depth [m]

Blob + TAB
Huh + TAB
0.08

0.06

0.04
Huh + Reitz&Diwakar
0.02

0
0 0.0005 0.001 We = 900
0.0015 0.002
Time [s]
© 2013 ANSYS, Inc. 8-20 Release 14.5
 Modeling Secondary Breakup
• “Statistical” approach
– Apply break-up process to “representative” particle “”
 Break-up process is applied to all “N” physical particles that are
represented by the “computational” particle
– If a particle (droplet) breaks-up then
• Adjust particle diameter and mass
• Adjust the number rate N
 mass of computational particle is conserved!
– Track single “representative” particle only

Break-up occurs

D1= 1, m1, N1 D2= 0.5  m2 = 1/8*m1, N2 = 8*N1

mt = m1*N1 mt = m2*N2= m1*N1

© 2013 ANSYS, Inc. 8-21 Release 14.5

 Secondary Breakup Models in CFX
• In ANSYS CFX, the following secondary break-up models are available:
– Hard coded models:
• Reitz & Diwakar (1987) Primary Break-up Secondary Break-up

• Schmehl Model (Schmehl, 2000) Dense Spray Dilute Spray

• TAB Model (O’Rourke, 1987)

• ETAB & CAB Model (Tanner, 1997&2003) Dispersed Flow
Liquid Core
– All model constants are accessible through CCL
– Alternative break-up models can be implemented with
Particle User Fortran
• The user implemented break-up models can depend on
– Particle and fluid variables
– Track information (e.g.: traveling time or distance)
– Randomness if required

© 2013 ANSYS, Inc. 8-22 Release 14.5

 Secondary Breakup Models GUI

• Domain/Fluid Pair Models

– Select particle breakup

model from GUI

– Access to model specific

constants

© 2013 ANSYS, Inc. 8-23 Release 14.5

 Model of Reitz & Diwakar

• Reitz and Diwakar

– Droplet break-up due to aerodynamic forces occurs in two modes:
• “Bag break-up”
– Non-uniform pressure field around droplet causes it to expand and
eventually disintegrate when surface tension forces are overcome
• “Stripping break-up”
– Liquid is sheared (stripped) from the droplet surface

– Droplet diameter change is described by:

dD p D p  D p , stable

dt b

© 2013 ANSYS, Inc. 8-24 Release 14.5

 Model of Reitz & Diwakar

Bag Stripping
Break-up Break-up

Vs2 D p We
 Cs1
Criterion We   2Cbl
p Re p

Characteristic  p D 3p Dp p
 b  Cb 2  b  Cs 2
Breakup-up time 2 p 2Vs 

Product 12 p  p2
D p , stable  D p , stable 
Droplet radius
V s
2
 2Vs3

© 2013 ANSYS, Inc. 8-25 Release 14.5

 Break-up Model of Schmehl
• Schmehl, 2003
– Two-step break-up process:
• Particle is deformed from sphere shape (‘0’) into disc shape

t:’0’ti

• Further distortion of droplet until final destruction

t:titi+tb

© 2013 ANSYS, Inc. 8-26 Release 14.5

 Break-up Model of Schmehl (cont.)

• Characteristic time:

D0 p
t 
*

Vslip g
ti
• Initial deformation of droplet into disc shape: *
 1.6
t

• Characteristic break-up time:

 6(We  12) 0.25 12  We  18

tb 
*
  2.45(We  12) 0.25
18  We  45
t 14.1(We  12) 0.25
 45  We  351
© 2013 ANSYS, Inc. 8-27 Release 14.5
 Break-up Model of Schmehl
• Mass mean diameter is calculated from
0.25
D0.5  1.2 * D0 * (1.5  Oh 0.2  Wecorr )
We m
Wecorr  Oh = Ohnesorge Number =
1  1.077Oh1.6 P  D P

• Product droplet velocity

2.5  D0
Vnorm 
2(tb  ti )

– D0 : Droplet diameter at beginning of deformation

© 2013 ANSYS, Inc. 8-28 Release 14.5

 Break-up Model of Schmehl
• Modified drag coefficient is used to account for droplet deformation due to
aerodynamic forces

21 6
cD  0.28  
Re P Re P
 We(0.2319  0.1579 log Re P  0.0471 log 2 Re P
 0.0042 log 3 Re P )
– Valid in the range of: 5 < ReP < 2000

© 2013 ANSYS, Inc. 8-29 Release 14.5

 Taylor Analogy Break-up Model

• TAB (Taylor Analogy Break-up)

– Drop deformation is described by forced, damped, harmonic oscillator

mx  F  bx  cx
Air flow
– F: force term (aerodynamic drag)
-x x
. force (surface tension)
– cx: restoring
– bx: damping term (viscosity)
Distorted
particle

Undistorted
particle
Distorted
particle

© 2013 ANSYS, Inc. 8-30 Release 14.5

 Taylor Analogy Break-up Model
• Modeling the break-up process:
• Solve equation (1) for oscillating droplet
.
– Deformation x and deformation velocity x known

– Droplet distortion x > xCrit? No

Yes• Determine child droplet size, rChild
• Determine droplet normal velocity component, VN
• Determine droplet deflection,  .
• Assume undistorted child droplet (x=x=0)

© 2013 ANSYS, Inc. 8-31 Release 14.5

 Taylor Analogy Break-up Model
• O’Rourke, 1987
• Break-up criterion:
– Break-up occurs when x > a/2
– a is the droplet radius at Wecrit = 6

• Child droplet size determined from:

rParent
rChild  2
 P rParent 3
 x   6K  5 
1  0.4 K     
  Cbr rParent   120 

© 2013 ANSYS, Inc. 8-32 Release 14.5

 Taylor Analogy Break-up Model
• Child droplets inherit a normal velocity component
VN  CV x
.
– x : Velocity of the equator of the parent droplet at the time of breakup

• Spray angle predicted from known particle velocity VP and normal velocity
component VN

• Shortcomings:
– Product droplet size often underestimated and therefore spray penetration too
short

© 2013 ANSYS, Inc. 8-33 Release 14.5

 Enhanced TAB
• ETAB (Enhanced Taylor Analogy Break-up)
• Tanner, 1997
• Extension of standard TAB model
. x0
– Initial deformation rate
• Break-up criterion same as for TAB model
• Product droplet size determined from:
rChild
 e  K br t
rParent
– with:  k1 We  Wet and  (frequency)
K br  
k 2 We We  Wet
40  Wet  100

© 2013 ANSYS, Inc. 8-34 Release 14.5

 Enhanced Taylor Analogy Breakup
• Child droplets inherit a normal velocity component:

VN  Ax

– A is determined from energy balance consideration

– x Velocity of the equator of the parent droplet at the time of
breakup

• Spray angle predicted from know particle velocity VP and

normal velocity component VN

© 2013 ANSYS, Inc. 8-35 Release 14.5

 CAB Break-up Model
• CAB (Cascade Atomization and Drop Break-up)
• Tanner, 2003
• Extension of standard ETAB model
• Break-up criterion same as for TAB model
• Product droplet size determined from:

rChild
 e  K bu t
rParent

k1 if 5  We  80

K bu  k 2 We if 80  We  350

k
 3  We 3/ 4
if 350  We
© 2013 ANSYS, Inc. 8-36 Release 14.5
 Secondary Droplet Breakup
0.1

• CAB model shows good agreement for

0.08
“medium” range injection Weber number (Weinj

Penetration Depth [m]

~103) 0.06

• Fails to predict spray formation at conditions 0.04

corresponding to “low” (~10) and “high” (~104) Experiment
injection Weber numbers. 0.02 We = 900
We = 2000
We = 3100
• “Bag Breakup Factor” was found to be major 0
0 0.001 0.002 0.003 0.004
constant that influences spray behavior Time [s]

• Contrary to the standard CAB model it is 0.08

Experiment
suggested that this constant should depend on We = 20000

Penetration Depth [m]

the injection Weber number 0.06

0.04

0.02

0
0 0.0005 0.001 0.0015 0.002 0.0025
Time [s]
© 2013 ANSYS, Inc. 8-37 Release 14.5
 Extension of CAB Droplet Breakup Model
0

Experiment
CAB
MCAB
0.004
Z [m]

High Pressure Diesel Injection

Weinj = 20000
0.08
0.008
Experiment
CAB
MCAB

Penetration Depth [m]

0.06
0 0.004 0.008 0.012
X [m]
0.04
Droplet in Crossflow
Weinj = 58
0.02

0
0 0.0005 0.001 0.0015 0.002 0.0025
Time [s]
© 2013 ANSYS, Inc. 8-38 Release 14.5
 Extension of CAB Droplet Breakup Model
Droplet diameters Droplet velocities
CAB 25 15
E xperiment
Experimental

D istribution function [% ]
Distribution function [%]
20 C AB
CAB
10
15

10
5

0 0
0 1E-05 2E-05 3E-05 4E-05 0 20 40 60 80
25 Diameter [m] V elocity [m/s]

MCAB Experimental
15
Distribution function [%]

20 E xperiment
MCAB

Distribution function [% ]
M C AB

15 10

10
5

0
0 0 20 40 60 80
0 1E-05 2E-05 3E-05 4E-05 V elocity [m/s]
Diameter [m]
© 2013 ANSYS, Inc. 8-39 Release 14.5
 Extension of CAB Droplet Breakup Model

• Model extension can be turned on/off via the following Expert Parameter

pt cab breakup improved = T/F

• For more details see:

– “Extension and validation of the
CAB droplet breakup model to a
wide Weber number range”.
In proceeding of ILASS Conference,
Como Lake, Italy, 2008,
Paper ID ILASS08-4-5.

© 2013 ANSYS, Inc. 8-40 Release 14.5

 Secondary Droplet Break-up Validation
Case 1 Case 2 Case 3

Gas parameters

Gas type N2

Temperature [K] 300

Pressure [MPa] 1.1 3.0 5.0

Fuel Properties

Fuel type C12H26

Density [kg/m3] 840

Surface tension [kg/s2] 0.0205

Spray parameters

Initial Temperature [K] 300

Injection Velocity [m/s] 102 90 86

Droplet diameter [mm] 0.3

Injection rates Single t=2.5 ms

pulse,
Initial Spray angle 7.5
[deg]
© 2013 ANSYS, Inc. 8-41 Release 14.5
 Validation: Spray Penetration – Case 1

t = 0.2 ms

t = 0.4 ms

t = 0.6 ms

t = 0.8 ms

t = 1.2 ms

© 2013 ANSYS, Inc. 8-42 Release 14.5

 Penetration Diagnostics
• Used to characterize the penetration SR
characteristics of a spray

• Available methods:
– Radial penetration depth (SR)
SN
– Penetration depth along a specified direction
(SP)
– Penetration depth normal to a specified
SP
direction (SN) PARTICLE PENETRATION: Origin 99
Option = Distance to Penetration Origin
– (Half) spray angle  Penetration Origin = 0.01 [m], 0.5 [m], 0.01 [m]
Mass Fraction Within Penetration Depth = 0.99
Particles List = Water
– Always calculated for a specified END
PARTICLE PENETRATION: Along Penetration Direction 99
mass fraction in the spray Option = Distance Along Penetration Direction
Penetration Origin = 0.01 [m], 0.5 [m], 0.01 [m]
(typically 99%) Penetration Direction X Component = 1.0
Penetration Direction Y Component = 0.0
Penetration Direction Z Component = 1.0
– Transient only! Mass Fraction Within Penetration Depth = 0.99
Particles List = Water
END
• In solver monitor and out-file PARTICLE PENETRATION: Normal to Penetration Direction 99
Option = Distance Normal to Penetration Direction
Penetration Origin = 0.01 [m], 0.5 [m], 0.01 [m]

• Beta-feature (CCL only)

Penetration Direction X Component = 1.0
Penetration Direction Y Component = 0.0
Penetration Direction Z Component = 1.0
Mass Fraction Within Penetration Depth = 0.99
Particles List = Water
END

© 2013 ANSYS, Inc. 8-43 Release 14.5

 Validation: Spray Penetration Case 1 & 3
Case 1

Case 3

© 2013 ANSYS, Inc. 8-44 Release 14.5

 Particle Mass Transfer
• Mass transfer from liquid particles to the continuous phase are
modeled in the spray drier and oil combustion models
• Mass transfer of volatile species from solid particles to the
continuous phase are modeled in the spray drier model
• Droplets or particles can consist of a single material or they
may contain multiple species

© 2013 ANSYS, Inc. 8-45 Release 14.5

 Multicomponent Particles
• Particles can consist of n different components
• Components can
– Be inert (Default)

– Evaporate into fluid phase

• Simple mass transfer model
– Diffusion & Convection
• Liquid evaporation model
– Diffusion & Convection & Boiling

– React (e.g.: Coal combustion)

– Condense on particle (User Fortran)

© 2013 ANSYS, Inc. 8-46 Release 14.5

 Multicomponent Particles
• Restrictions for particle component materials:
– Density must be constant
– Heat capacity must be constant or specified using NASA polynomials

• Track file contains:

– Total particle mass
– Mass fractions of all particle components

© 2013 ANSYS, Inc. 8-47 Release 14.5

 Particle Heat Transfer

• Particles exchange energy with the surrounding fluid

• Energy exchange can happen due to:
– Convection
– Radiation (particularly important in oil and coal combustion)
– Mass transfer

© 2013 ANSYS, Inc. 8-48 Release 14.5

 Particle Heat Transfer - Options

Available Options:

• Ranz Marshall
– Convection
• Radiation
• None
• User Defined

© 2013 ANSYS, Inc. 8-49 Release 14.5

 Particle Heat Transfer Terms
• Particle heat transfer terms are
selected from within the
Fluid Pairs section of the
Domain panel

• Convective heat transfer

(can be switched off with
Option = None)
• User defined
(Particle User Routine)

© 2013 ANSYS, Inc. 8-50 Release 14.5

 Convective Heat Transfer

• Convective heat transfer

QC  d P Nu (TF  TP )

• Nusselt Number, Nu

Nu  2  0.6 Re1/ 2 ( c p /  )1/ 3 (Sphere)

diffusion convection

© 2013 ANSYS, Inc. 8-51 Release 14.5

 Radiative Heat Transfer
• Radiative heat transfer
1
QR   Pd P2 ( I   n 2TP4 )
4
• Particles receive radiation

I
• Particles emit radiation

 TP4

• Particle contributions to radiation can

be fed back into radiation field Note: No
radiation within particles.
© 2013 ANSYS, Inc. 8-52 Release 14.5
 Energy Transfer via Mass Transfer

• Energy transfer associated with mass transfer

and associated enthalpy

dmC
QM   VC
dt

• Particle components evaporate at a rate

of dmC dt
T

• Latent heat of evaporation, V:

heat necessary to evaporate a substance
Liquid+
without temperature change (but phase Liquid Gas
Gas
change!)
Heat
V
© 2013 ANSYS, Inc. 8-53 Release 14.5
 Latent Heat of Evaporation
• Latent heat of evaporation V
– Specified in Component Pairs section of
CFX-5 Pre

•Option = Automatic
– Thermodynamic properties are consistent
(i.e. in the same ‘Material Group’ or use
consistent reference enthalpies)

•Option = Latent Heat

– Reference temperature and latent heat at
reference temperature must be specified
– Use if thermodynamic properties for the gas
and liquid phase components are not consistent

© 2013 ANSYS, Inc. 8-54 Release 14.5

 Particle Mass Transfer

• Particle components can exchange mass with the surrounding fluid

• Mass exchange can happen due to

– Diffusion
– Convection
– Boiling
– Reaction

© 2013 ANSYS, Inc. 8-55 Release 14.5

 Particle Mass Transfer Options
Available options:
• Ranz Marshall
– Diffusion/Convection
(Simple Mass Transfer Model)

• Liquid Evaporation Model

– Diffusion/Convection and Boiling
(Spray Drier Model)

• None
– No mass transfer
– Allows the usage of user defined
mass transfer terms done with
particle user fortran routines

© 2013 ANSYS, Inc. 8-56 Release 14.5

 Simple Particle Mass Transfer Model

• Mass transfer of a component due to molecular diffusion and

forced convection

dmC
 (C P ,C  C F ,C )  d P D Sh( R  YP ,C  YF ,C )
dt
with: D: Dynamic diffusivity
Y: Component mass fractions in particle and fluid
R: Mass fraction equilibrium ratio

• Sherwood number correlation:

Sh  2  0.6 Re1/ 2 ( / D)1/ 3
molecular forced
diffusion convection

© 2013 ANSYS, Inc. 8-57 Release 14.5

 Interfacial Equilibrium Model

• Control the value of the Mass

Fraction Equilibrium Ratio
• Mass fractions are discontinuous
at particle surfaces, and the mass
fraction in the continuous phase at
the particle surface is this ratio
multiplied by the mass fraction in
the particle
• If not defined, then YP,C is
computed from:
YP,C = component mass/particle mass

© 2013 ANSYS, Inc. 8-58 Release 14.5

 Liquid Evaporation Model

• Synonym: “Spray drier model”

• Main difference to simple mass transfer model:
– Particle can boil
• Mass transfers rates change, if

– TP<Tboil
dmC
 Diffusion and/or Convection
dt
– TP  Tboil
dmC
 Heat transfer limited
dt

© 2013 ANSYS, Inc. 8-59 Release 14.5

 Liquid Evaporation Model

• Particle is below boiling point

dmC WP ,C  1  X P ,C 
 dDSh log 

dt WF ,C  1  X F , C 

• Particle above boiling point

dmC  QC  QR 

dt V

© 2013 ANSYS, Inc. 8-60 Release 14.5

 Liquid Evaporation Model - Boiling
• Boiling point TP,boil = f(p,Material)
• TP,boil is determined by evaluating an Antoine equation

 B 
pvap  exp A  
 TP  C 
• The Antoine equation is a parameter fit to experimental vapor pressures
– A, B, and C are "Antoine coefficients"
– Antoine coefficients are specified in ANSYS CFX via a “Homogeneous
Binary Mixture” (see next slide)

• Particle boils if pvap > pambient

© 2013 ANSYS, Inc. 8-61 Release 14.5

 Homogeneous Binary Mixture
• Homogenous Binary Mixture
– Define link between particle material and
corresponding fluid material
– Define Antoine coefficients for evaporation
process on the Saturation
Properties tab under the Pressure option

© 2013 ANSYS, Inc. 8-62 Release 14.5

 Light Oil Modification

• Extension of standard
Liquid Evaporation Model
– Evaluation of Re, Nu and Sh
numbers in heat/mass transfer
correlations is based on ‘average’
properties within boundary layer of the
particle

Particle Fluid
  0.0 if pVap  pambient
  0.5 if pVap  pambient
Vol   Fl (1   )  Vol  Fl
pVap
  0.5
Boundary Layer pambient

© 2013 ANSYS, Inc. 8-63 Release 14.5

 Fates for Evaporating Particles
• Particles are tracked until they fall below a minimum diameter:
– REAL expert parameter: PT MINIMUM DIAMETER
+--------------------------------------------------------------------+
| Particle Fate Diagnostics |
+--------------------------------------------------------------------+
| Particle type | Fate type Particles |
+--------------------------------------------------------------------+
| JetA Liquid | Entered domain : 4 |
| | Fell below minimum diameter : 4 |
+--------------------------------------------------------------------+
• Source term diagnostics:

+--------------------------------------------------------------------+
| Particle Equation | Total source and source change rates |
+--------------------------------------------------------------------+
| | Equation Source Rate |
+--------------------------------------------------------------------+
| Domain: Domain 1 |
+--------------------------------------------------------------------+
| JetA Liquid | Mass-JetA 1.192E-04 0.0014 |
+--------------------------------------------------------------------+
| JetA Liquid | x-Mom 1.189E-04 0.0014 |
| | y-Mom 0.000E+00 0.0000 |
| | z-Mom 0.000E+00 0.0000 |
+--------------------------------------------------------------------+
| JetA Liquid | Energy 2.162E+02 0.0012 |
+--------------------------------------------------------------------+

© 2013 ANSYS, Inc. 8-64 Release 14.5

 Under Relaxation Factors

• Under relax particles sources onto fluid phase

Srcn  1    Srco   Srcn

– Momentum
• Velocity Under Relaxation Factor (Default=0.75)
– Energy
• Energy Under Relaxation Factor (Default=0.75)
– Mass Transfer
• Mass Under Relaxation Factor (Default=0.75)

© 2013 ANSYS, Inc. 8-65 Release 14.5

 Erosion Modeling
Erosion is

The group of natural processes by which material is worn away from a

surface

VP


• Erosion rate, E, is a function of

– impact angle and velocity, material combination, temperature, particle
shape, …
– Dimension: [mass removed/mass impacted]
• Erosion models
– are ‘curve fits’ to experimentally obtained erosion rates
– differ in number of variables taken into account to determine erosion rate, E

© 2013 ANSYS, Inc. 8-66 Release 14.5

 Erosion Modeling Overview

Finnie Tabakoff & Grant

E  f (VP ,  ) E  f (VP ,  ) and material combination

n 2
V  V 
E   P  f ( ) E  f ( ) P  cos 2  [1  RT2 ]  f (VPN )
 V0   V1 
1   90 
f ( )  sin 2    71.5 f ( )  1  k 2 k12   
  0 
3

f ( )  sin( 2 )  3 cos 2  71.5    90
V
RT  1  P sin 
V3
4
V 
f (VPN )   P sin  
 V2 
Constants: V0 ,n Constants: V1 , V2 ,V3 , k12 ,  0

© 2013 ANSYS, Inc. 8-67 Release 14.5

 Erosion Modeling – Setup
• Access via the Domain Fluid Values tab
• Available erosion models:
– Finnie
– Tabakoff & Grant
– User Defined (User FORTRAN)
• Model handling
– Domain or patch-by-patch basis
• Each surface patch can have a
different erosion model
• Possible to limit erosion prediction to
user specified surfaces
– User access to all erosion model constants via
CCL
– Erosion rate density is written to res file for post-processing
– Tabakoff & Grant: Material combinations
• Quartz/Aluminium (default), Quartz/Steel, Coal/Steel

© 2013 ANSYS, Inc. 8-68 Release 14.5

 Wall Interaction and Wall Films
• Particle-Wall Interaction Model
– This model describes how particles interact with film covered walls and under
which conditions a wall film is formed
 Prerequisite for wall film model

• Droplet-wall interaction is complex and not all aspects are well understood.
– Dimensional analysis shows that droplet-wall interaction depends on:
– Particle quantities (Weber Number), existence of a wall film, wall roughness,
wall temperature (and much more)

© 2013 ANSYS, Inc. 8-69 Release 14.5

 Particle Wall Interaction

• BC Panel  Fluid Values

• Set “Wall Interaction” option to ‘Wall
Film”
– This activates existance of wall film on that
patch
• Select Wall Interaction Model
– Elsaesser
– Input: Wall Material
– User Defined
– Input: User Routine and Argument list
to User Fortran code
– Stick to Wall (CCL only)
– No further input required

© 2013 ANSYS, Inc. 8-70 Release 14.5

 Wall Interaction for the Elsaesser Model
Cold-Wetting Hot-Wetting Hot-Nonwetting
WeP
Splashing Splashing-
Boiling- Breakup
Wall-Breakup

Boiling- Rebound
Spreading Wall-Breakup with Breakup

Boiling-
Rebound
Breakup
Bouncing
TPa TPr
Twall

Tpa and Tra are the so called ‘pure adhesion’ and ‘pure rebound’ temperatures;
both depend on combination of wall/particle material
© 2013 ANSYS, Inc. 8-71 Release 14.5
 Elsaesser Model Wall Roughness

• Wall roughness is idealized:

Rt: Roughness height
Rt : Roughness angle,
  = f()

• Particle reflection depends on relative size of particle wrt. wall

roughness height (=Rt/dp)

 < 0.25  = 0.5 >1

© 2013 ANSYS, Inc. 8-72 Release 14.5

 Notes on the Wall Interaction Models
• Elsässer Model
– Accounts for wall temperature effects, wall roughness and particle-wall material
combination, …
– Targeted towards IC-E applications (~ Gasoline injection)

• Stick to Wall
– Simplest possible model: all particles that hit a wall become part of the wall film

• User Fortran Interface

– Option for user supplied FORTRAN routine

© 2013 ANSYS, Inc. 8-73 Release 14.5

 Spin-off of Wall Interaction Extension
• Child droplet generation model
– Parent droplet can create more than one
child
• Generic access to “wall only” variables in
UF
– Wall temperature, wall roughness, wall
velocity, etc…

© 2013 ANSYS, Inc. 8-74 Release 14.5

 Particle Wall Film
• Major physical phenomena

Splashing
Evaporation Convection
External Forces Separation
Impinging

Conduction

• Film movement due to external forces is neglected in

ANSYS CFX
– Film can still move if on a moving wall

© 2013 ANSYS, Inc. 8-75 Release 14.5

 Wall Film Modeling
• Modeling Approach
– Wall film is modeled using a Lagrangian approach
– i.e.: Wall film made of a special type of particles
 “Wall particles”

• Assumptions
– Thin film approach (no displacement effect)
– Neglect influence of film on fluid drag
– No film movement due to external forces
 Quasi Static Wall Film

© 2013 ANSYS, Inc. 8-76 Release 14.5

 Wall Film Example: Evaporating Droplets

Water droplets (Tp = 293 K) hit a heated

wall
Assumptions:
Twall = 350 [K]
Qconvective • Droplets stick to wall, i.e. no relative
movement between particles and wall
• Energy is transferred from
wall/surrounding to film
Qconduct

• Film evaporates into ambient

© 2013 ANSYS, Inc. 8-77 Release 14.5

 Wall Film Example: Evaporating Droplets

© 2013 ANSYS, Inc. 8-78 Release 14.5

 Wall Film Example: Evaporating Droplets

© 2013 ANSYS, Inc. 8-79 Release 14.5

 Wall Film Example: Evaporating Droplets

© 2013 ANSYS, Inc. 8-80 Release 14.5

 Wall Film Example: Evaporating Droplets

© 2013 ANSYS, Inc. 8-81 Release 14.5

 Wall Film Example: Evaporating Droplets

© 2013 ANSYS, Inc. 8-82 Release 14.5

 Wall Film Output and Restrictions
• Output Variables:
– Wall film thickness
– Averaged wall film temperature
– Total wall film mass (via TPD)
– New particle fate (‘Wall particles”)
• Restrictions:
– Wall films of different particle types do not mix.
– Wall films of different particle types cannot coexist on the same boundary patch
– Wall film model only available for transient and turbulent simulations

© 2013 ANSYS, Inc. 8-83 Release 14.5

© 2013 ANSYS, Inc. 8-84 Release 14.5