ANSYS Fluent Multiphase Tips and

Tricks
Outline
• VOF model setup
• Use of Open Channel conditions with VOF
• VOF numerical stabilization
• Free surface applications of the VOF model
• Best practices for modeling dispersed flows
• Overview of phase change models
• Boiling and thermal phase change
• Evaporation-condensation modeling
• Solutions for modeling cavitation
 VOF
• Model two or more immiscible fluids by solving a single set of momentum equations
and tracking the volume fraction of each of the fluids throughout the domain
• Typical applications are: jet breakup, large bubbles in a liquid, dam break, tracking of
any liquid-gas interface
• Following restrictions apply to the VOF model in ANSYS Fluent
– Must use pressure-based solver. Cannot use density-based solver
– Cannot model Streamwise periodic flow
– Cannot use second-order implicit time-stepping formulation with the VOF explicit scheme
– No void regions. All control volumes must be filled with single fluid phase or a combination of phases
– Only one phase can be defined as a compressible ideal gas. Can use UDF to define compressible liquids
– When tracking particles in parallel, the DPM model cannot be used with the VOF model if the shared
memory option is enabled (using the message passing option, when running in parallel, enables the
compatibility of all multiphase flow models with the DPM model)

3 © 2015 ANSYS, Inc. June 12, 2018

 Initialize a Non-Flat Free Surface in ANSYS Fluent
Search Keywords: initialize non-flat free
surface

Customer Portal KM ID: 2041321

This solution describes three ways to start a simulation with non-flat free surfaces:
• Using Open Channel Wave Model to initialize with a wavy free surface
• Using Custom Field Functions to patch a non-flat free surface
• Using UDFs to initialize with a non-flat free surface

4 © 2015 ANSYS, Inc. June 12, 2018

 Initialize a Non-Flat Free Surface in ANSYS Fluent Continued…
• Using Open Channel Wave Model to initialize with a wavy free surface

5 © 2015 ANSYS, Inc. June 12, 2018

 Initialize a Non-Flat Free Surface in ANSYS Fluent Continued…
• Using Custom Field Functions to patch a non-flat free surface
– Define CFF 1

– Use adaption registers to mark the cells

– Use Patch, select phase, value and register

3
2

6 © 2015 ANSYS, Inc. June 12, 2018

 Initialize a Non-Flat Free Surface in ANSYS Fluent Continued…
• Using UDFs to initialize with a non-flat
free surface
– DEFINE_INIT
– Important to specify the volume fraction of all
phases in the UDF
– Loop over all phases (subdomains) in the
mixture domain

7 © 2015 ANSYS, Inc. June 12, 2018

 Stabilizing VOF Simulation
Search Keywords: Stabilize VOF Simulation

Customer Portal KM ID: 2039870, 2039871

Below are some of the settings that can help stabilize the solution:
1. Use variable time stepping with following settings
– Global Courant Number = 2 to 0.1, Minimum Step Change Factor = 0.05, Maximum Step Change Factor
= 1.1
2. Lower under-relaxation factor to 0.2-0.3 for momentum
3. Symmetric vs Anisotropic drag law: For zero initial velocities or sharp interface,
anisotropic drag is recommend over Symmetric drag
– Anisotropic has additional treatments for stability

8 © 2015 ANSYS, Inc. June 12, 2018

 Stabilizing VOF Simulation Continued…
4. Running first time step is very crucial from solution stability perspective. In most of the
cases, stability issues are faced in first time-step itself
– Use realistic initial conditions, and lower CFL.
– By default, for first time-step, FLUENT solves VOF equation at every iteration. Though default setting
will work for most of the cases, in some cases it can lead to stability issues
– You can turn off the calculation of VOF equation at every iteration by changing following rpvar:
(rpsetvar 'mp/mfluid-first-time-step? #f)
5. Diffusive or mildly diffusive schemes work best with multi-fluid formulation
– Geo-Reconstruct with multi-fluid can be very stiff to converge, whereas compressive scheme is
recommended as it gives more solution stability

9 © 2015 ANSYS, Inc. June 12, 2018

 Stabilizing VOF Simulation Continued…
6. Additional solution stability suggestions for Geo-Reconstruct Schemes
– Lower under-relaxation for pressure and momentum = 0.2 - 0.3
– Using SIMPLE instead of PISO
– Replace Incompressible air with isothermal ideal gas
• Energy equation not being solved
• Allows variation in density for air
• Critical for closed domain with high rotation
– Use First Order for momentum equation

10 © 2015 ANSYS, Inc. June 12, 2018

 Speed Up VOF Simulation
Search Keywords: Speed Up VOF
Simulation

Customer Portal KM ID: 2037063

• Free surface flows are typically transient in nature and can become computationally
expensive due to time-step constraints
• Accuracy of VOF simulations depends on the discretization schemes chosen for volume
fraction equation and spatial gradients, besides other factors like mesh resolution
• Attached document provides guidelines that are expected to give the best trade-off
between accuracy and speedup for VOF simulations

11 © 2015 ANSYS, Inc. June 12, 2018

 Speed Up VOF Simulation Continued…
• Mesh Considerations for VOF Simulations
– Uniform mesh /Smooth transition in case of non uniform mesh
– Finer mesh for interface region
– Avoid highly skewed cells and high aspect ratio cells
– Same mesh type in the interface region
• For speed up with Explicit VOF
• For less numerical diffusion with Implicit VOF

12 © 2015 ANSYS, Inc. June 12, 2018

 Speed Up VOF Simulation Continued…
• Use Implicit VOF, Compressive and Bounded Second Order Time Discretization scheme
– This allows you to use a larger time step size
– Use higher URFs for pressure and momentum (up to 0.8)
– Try NITA

• If the solution is not accurate with Implicit VOF

– Check the solution with a finer mesh and a smaller time step size
– Use Explicit Compressive or Geo-Reconstruct

• Explicit VOF
– Use uniform mesh in the liquid interface regions
• Use Variable Time Stepping for non-uniform mesh in the interface region
– Try with different Courant calculation methods: Solve > set > vof-explicit-controls
– Try NITA

13 © 2015 ANSYS, Inc. June 12, 2018

 Liquid-gas interfaces are ubiquitous in industry and the environment

Nuclear cooling towers [1] Fluid catalytic cracking stations [2] Burning spray of
liquid ethanol [3]

[1] senseandsustainability.net
[2] Harman et al., J. Aerosol
Science, 1999
[3] seia.org
“The breaking of waves is a process that is ubiquitous [4] geiconsultants.com
over two-thirds of the surface of the globe… It leads to [5] KM 2044984
Water conveyance pipelines [4]
enhancement Breaking
of heat of an ocean
transfer and wave [5] of gas
especially Flow over weirs and spillways [6]
[6] commons.wikimedia.org
exchange between the atmosphere and the ocean…”
- Owen Phillips, J. Fluid Mech., 1985

14 © 2015 ANSYS, Inc. June 12, 2018

 It is numerically impractical to advance the free surface by directly
discretizing the governing equations
• Mass continuity equation (after incompressible assumption) given by

• Navier-Stokes equation describes momentum conservation:

• What happens at the free surface between the immiscible fluids?

Large differences in fluid properties (mainly the density) make it

numerically difficult to directly discretize and solve the continuity
equation

15 © 2015 ANSYS, Inc. June 12, 2018

 Different interface transport techniques
• MOVING GRIDS
• Accurate
• Topology changes more difficult to handle

• MARKER PARTICLES
• Accurate
• Topology changes more difficult to handle

• VOLUME OF FLUID
• Good volume conservation
• Automatic topology changes
• Formal accuracy limited

• LEVEL SET
• Accurate and simple
• Automatic topology changes Images reproduced from:
Scardovelli & Zaleski, ARFM, 1999
• Not inherently mass conserving

16 © 2015 ANSYS, Inc. June 12, 2018

 KM #1: Free surface modeling in Fluent
Search Text: multiphase free surface model

Customer Portal KM ID: 2046036

• KM: Do I need to use a multiphase model if I have a free surface?

Summary: Depending on the nature of the free surface and the objective(s) of the
simulation, the free surface may not need to be modeled

17 © 2015 ANSYS, Inc. June 12, 2018

 KM #1: Do I absolutely need a multiphase model?
• Benefits
− Less modeling required (we’re not
solving for the free surface position)

− Domain is smaller, requiring fewer

cells/no added resolution to capture
the interface

− More likely to be steady-state (even

with a nominally flat free surface,
small transient features tend to occur
which prevent the simulation from
converging)

18 © 2015 ANSYS, Inc. June 12, 2018

 KM #1: Examples that may not require including the free surface
• Mixing tank
− The interest is in the mixing caused by the blades
− In a well-designed (baffled) system where the blades are not adjacent to the liquid surface, the surface position
may be relatively level

• Water industry and river/lake/reservoir systems

− In many such systems the flow is relatively slow and little/no agitation is required
− Surface ripples have little to no effect on the overall flow, meshing to resolve the free surface is impractical
− the presence of weirs or hydraulic jumps (spill ways) will require the free surface to be included

• Multiphase separator
− Liquid surface is often relatively stable
− Interest is focused on where trace droplets/bubbles/particles go within the liquid
− If no DPM, the free surface may need to be modelled as a “degassing” type boundary condition
19
−© 2015
If DPM,
ANSYS, Inc.
the free surface may be a slip wall with DPM trap
June 12, 2018
 KM #2: Free surface modeling in Fluent
Search Text: mixing tank surface

Customer Portal KM ID: 2046975

• KM: Vortex prediction in a mixing tank

Summary: Uses the MRF and sliding mesh models to predict and compare the
interface topology in a mixing tank

20 © 2015 ANSYS, Inc. June 12, 2018

 KM #2: Mixing tank setup

21 © 2015 ANSYS, Inc. June 12, 2018

 KM #2: MRF/Sliding zones, models, and solution controls
• VOF for multiphase
• Reynolds stress model (7 eqn model) for turbulence

Solution methods Solution controls Zone with frame

motion specified

After running case with MRF,

change to sliding mesh with the
TUI command

/mesh/modify-zones/mrf-to-sl
iding-mesh 5

22 © 2015 ANSYS, Inc. June 12, 2018

 KM #2: Frame motion and mesh motion comparison

Frame motion Mesh motion

23 © 2015 ANSYS, Inc. June 12, 2018

 KM #3: Free surface modeling in Fluent
Search Text: free surface ship

Customer Portal KM ID: 2046972

• KM: Free surface ship hull simulation

Summary: Modeling free surface flow around a ship hull using the VOF method
and comparing with experimental data

24 © 2015 ANSYS, Inc. June 12, 2018

 KM #3: Outline of KM and schematic of setup
• Use VOF multiphase model of Fluent to solve open channel flow
• Set the boundary conditions for an open channel multiphase flow
• Use the SST k-ω turbulence model
• Run the simulation until a steady-state solution is obtained
• Post-processing and comparison with experimental data

Problem schematic

Fr = 0.267

25 © 2015 ANSYS, Inc. June 12, 2018

 KM #3: Overview of results Comparison of water level on the hull surface

Transient development of air-water interface to steady-state

Sarda, O. P. “Turbulent Flows Past Ship Hulls

– An Experimental and Computational
Study”, PhD thesis, University of Iowa, 1986.

Free surface wave elevation (in meters)

26 © 2015 ANSYS, Inc. June 12, 2018
 KM #4: Free surface modeling in Fluent
Search Text: surface area vof

Customer Portal KM ID: 2041527

• KM: UDF to obtain the VOF surface area

Summary: The surface area of the phase interface in a VOF simulation can be
obtained to compute useful interfacial quantities

27 © 2015 ANSYS, Inc. June 12, 2018

 KM #4: Interface area calculated from gradient of volume fraction

• This surface area can be used to calculate physical quantities, e.g.

– Heat transfer from free surface
– Total surface energy

28 © 2015 ANSYS, Inc. June 12, 2018

 KM #4: UDF to compute VOF surface area

Necessary for
storing VOF
gradients

29 © 2015 ANSYS, Inc. June 12, 2018

 KM #5: Free surface modeling in Fluent
Search Text: free surface unstable

Customer Portal KM ID: 2048334

• KM: Interfacial anti-diffusion with extremely high viscosity ratio leads to unstable solution
Summary: A couple tips are presented which help eliminate spikes in pressure and
velocity when simulating high viscosity ratio fluids with VOF
30 © 2015 ANSYS, Inc. June 12, 2018
 KM #5: Problem and solution
• Problem
– Why does using interfacial anti-diffusion in a high viscosity ratio injection simulation show unstable spikes in
the velocity and pressure field?

• Solution
– For faster convergence in VOF simulations, implicit formulation for volume fraction can be used, allowing for
larger time steps. The interfacial anti-diffusion option can be used for increasing interfacial accuracy for sharp
interfaces. With this setup, one may observe convergence difficulty, resulting in spikes in the velocity and
pressure fields.
– The interfacial anti-diffusion is largely responsible for the issues: while this helps preserve the sharpness of
the interface, it is known to make convergence more difficult, especially when using aggressively large time
step sizes.
– In such cases, it is found that running with explicit VOF, geo-reconstruct and DISABLING interfacial
anti-diffusion eliminates the spikes. This does require running with time steps smaller than preferred, but
increasing the viscosity of air helps to increase the time step size (without altering the inter-phase physics too
much).

31 © 2015 ANSYS, Inc. June 12, 2018

 KM #5: Spikes occurring in velocity and pressure fields

Rubber material
initially from inlet
to here
Rubber-like
material in

Rubber-like
material
and air out
32 © 2015 ANSYS, Inc. June 12, 2018
 KM #5: Spikes removed from velocity and pressure fields

Rubber material
initially from inlet
to here
Rubber-like
material in

Rubber-like
material
and air out
33 © 2015 ANSYS, Inc. June 12, 2018
 Multiphase Flows
Today’s meeting is part of our global “Meet your KMs” series. The goal behind these meetings is to showcase the
solutions that are available on our customer portal and to encourage their use, both internally and externally

Multiphase flow are typically classified as

• Free Surface flows (Modeled using VOF methods)
• Dispersed flows (Mixture or Eulerian approach)
Todays session deals with a few best practices for modeling Dispersed flows

We will be covering a few important guidelines/best practices from these solutions available in the customer
portal. For more details, you are requested to go through the corresponding solution documents in the portal

34 © 2015 ANSYS, Inc. June 12, 2018

 Solution – I
Best Practice Guidelines for modeling Bubble Columns
Search Keywords: Bubble Columns best practices

Customer Portal KM ID: #2039372

Bubble columns are widely used in the

chemical/process industry
This document summarizes the best practices
that have been effective in the modelling of
bubble columns or similar applications

35 © 2015 ANSYS, Inc. June 12, 2018

 Bubble Column Modeling

• Pre-Processing Requirements
• Choice of dispersed multiphase model
• DPM, Mixture and Euler-Euler model
• Modelling of interfacial forces
• Drag, Virtual mass, Lift, Wall Lubrication forces
• Treatment of the liquid free surface

36 © 2015 ANSYS, Inc. June 12, 2018

 Bubble Column Modeling
Pre-processing requirements

• A good quality mesh must be ensured

• Sufficient resolution in radial and axial direction
• Use of a hex mesh is recommended
• Cell skewness should be below 0.5
• Acceptable aspect ratio for the cells, especially near
the gas-liquid interface

37 © 2015 ANSYS, Inc. June 12, 2018

 Bubble Column Modeling
Multiphase Approach

• Euler-Euler model is recommended for simulating

bubble columns
• It is the most comprehensive of all dispersed phase
multiphase models
• For simulations constrained by computing resources,
mixture model may also be used

38 © 2015 ANSYS, Inc. June 12, 2018

 Bubble Column Modeling
Transient behavior
Instant (a) and Average (b) bubble volume fraction predicted
by CFD model
• For hydrodynamic time scales, flow in bubble
columns in unsteady, and three dimensional
• Bubble plume oscillation is of the order of a few
Hertz, while turbulent fluctuations are of the order
of a few 100-1000 Hertz
• Instantaneous flow field is not symmetric
• To get reproducible symmetric flow fields, both
experimental and simulated data must be time
averaged for sufficient periods of time CFD Modeling of Slurry Bubble Column Reactors for Fisher-Tropsch Synthesis
Andrey A. Troshko, Franz Zdravistch, Chemical Engineering Science 64 (2009)
892-903

39 © 2015 ANSYS, Inc. June 12, 2018

 Bubble Column Modeling
Modelling interaction between phases
• Drag between the liquid phase and gaseous phase is the most important interaction
term between the phases
• Fluent offers the ability to model the following additional interaction terms which
normally are important to include in bubble columns
• Lift, Virtual mass ,Wall Lubrication, Turbulence Dispersion & Turbulence Interaction (Bubble
Induced Turbulence)
Interfacial forces Comments
Drag Forces the bubbles in the flow direction

Lift Forces the bubbles to move perpendicular to flow direction

Wall lubrication Pushes the bubbles slightly away from the wall

Turbulent Tries to homogenize the volume fraction gradients..(Turbulent

dispersion diffusion)

Virtual mass Force acting on continuous phase due to dispersed phase acceleration

40 © 2015 ANSYS, Inc. June 12, 2018

 Effect of Interface forces

Profiles of time averaged simulated and measured axial liquid velocity for different interfacial
forces, at a height of 0.25m and a depth of 0.075m.
D: drag force, L: lift force, VM: virtual mass force (from Deen, 2001).

http://vbn.aau.dk/files/19821395/phd_thesis_hansen.pdf
The above picture shows that modeling drag alone results in incorrect velocity profiles
for this bubble column. Including lift force gives an improved prediction of velocities

41 © 2015 ANSYS, Inc. June 12, 2018

 Modeling the free surface

• The free surface at the top of the column can be Free Surface

modelled implicitly by including the freeboard

region on top of the column and specifying pressure
outlet boundary condition
• The free surface at the top of the column can also be
approximated by the use of a degassing boundary
condition
• The absence of large gradients makes the numerics more
stable in this case
• Degassing boundary condition is recommended for Bubble
column simulations

42 © 2015 ANSYS, Inc. June 12, 2018

 Recommended Settings
Models
Multiphase Euler – Euler (Implicit)
Turbulence SST K-omega (or RNG K-epsilon)

Phases Setup
Phase Specification Primary Phase: Liquid
Secondary Phase: Gas
Phase Interaction Drag: Grace Drag Force (or Universal Drag law)
Lift: Tomiyama lift force
Wall Lubrication: Antal et al (default coeff.)
Turbulent Dispersion Burns et al. (cd=0.8)
Turbulent Interaction Sato Model (default coeff.)
Surface Tension Coeff.: Specified

43 © 2015 ANSYS, Inc. June 12, 2018

 Recommended Settings
Solution Methods
Pres.-Vel. Phase Coupled SIMPLE
Coupling
Spatial Gradient: Least Squared Cell Based
Discretization Momentum: QUICK
Volume Fraction: QUICK
TKE: 1st Order Upwind
Specific Diss. Rate:1st Order Upwind

Transient Bounded 2nd Order Implicit

Formulation

44 © 2015 ANSYS, Inc. June 12, 2018

 Solution – II
Applicability of Phase Change Models in Fluent
Search Keywords: Phase Change models
applicability
Customer Portal KM ID: 2041749

This solution helps in guiding the users to

understand the applicability of the models and
their expected behavior for a particular
application

45 © 2015 ANSYS, Inc. June 12, 2018

 Phase Change Models in ANSYS FLUENT
In ANSYS Fluent the following are available for modeling phase
change
• Boiling models
• RPI
• Non equilibrium
• Critical Heat Flux
• Evaporation Condensation
• Thermal Phase Change
• Lee
• In addition there are “Cavitation” and “Wet steam” models
also available
• Cavitation model is suited for predicting cavitation
• Wet steam model is suited for applications such as rapid
expansion of steam which causes condensation
• This document discusses the applicability of Wall boiling
models and Evaporation Condensation models alone

46 © 2015 ANSYS, Inc. June 12, 2018

 Phase Change Models in ANSYS FLUENT
• Wall boiling models
– Wall boiling model is applicable for subcooled nucleate boiling in vertical channels. Critical
Heat
– RPI Boiling model Flux Minimum
Heat Flux

Heat Flux
• Applicable to subcooled nucleate boiling
– Non-equilibrium Boiling

Saturated
• Extension of RPI to take care of saturated boiling Subcooled
Transitio Stable
– Critical Heat Flux Nucleate
nal or
boiling Film
Unstable
• Extension of RPI to take care of boiling crisis Single boiling
Phase
• Lee Model Wall Superheat (Twall - Tsat)
– Lee Model is a simplified approach to model volumetric phase change where user needs to tune the
coefficients to match the results with the known experimental results. Even though this model can be used
with the Eulerian model, it is recommended to be used with the Mixture multiphase model.

• Thermal Phase Change Model

– Thermal Phase Change Model is also a volumetric phase change model. However, unlike Lee Model, there is
no need to tune the coefficients for mass transfer. Mass transfer is governed by the chosen heat transfer
mechanism. This model is available with the Eulerian multiphase model.

47 © 2015 ANSYS, Inc. June 12, 2018

 Sample Case
A sample case where liquid at saturation conditions enters an
annular pipe and flows vertically upwards. The inner surface of the
pipe is heated which causes evaporation of the fluid

RPI wall boiling model is suited for such conditions.

However for explanation purposes, the same case is run using

Thermal Phase model, and Lee model to explain the comparative
behavior.
This document does not explain the theory behind each of the
models and users are requested to refer to documentation section
for details.
Please note that this is not a validation study or a best
practice document.

48 © 2015 ANSYS, Inc. June 12, 2018

 Sample Case
Annulus geometry
•ID (mm): 141
•OD (mm) : 182
•Height (mm) : 200
•Inside pipe surface area (m2) = πx0.141x 0.2
= 0.0885

•Heat flux provided on wall (w/m2) = 1027500

•Heat supplied (w) = 1027500 x 0.0885 = 91029.2

Heat balance
Let m be mass of steam generated
M x Latent heat = heat supplied
M x (1262.3x 103)= 91029.2
Mass of steam = 0.072 kg/s

Theoretically maximum mass flow rate of steam at outlet = 0.072 Kg/s

49 © 2015 ANSYS, Inc. June 12, 2018

 RPI Boiling Model

Report -> Fluxes

Overall mass balance

Contours of Volume
fraction of Steam
Phase Interaction: Steam flow rate at outlet
• Drag: Ishii
• Lift: Moraga
• Heat transfer: Ranz Marshall

50 © 2015 ANSYS, Inc. June 12, 2018

 Thermal Phase Change Model

Report -> Fluxes

Overall mass balance

Contours of Volume
fraction of Steam

Steam flow rate at outlet

51 © 2015 ANSYS, Inc. June 12, 2018

 Lee Model

Report -> Fluxes

Contours of Volume
fraction of Steam Overall mass balance

Steam flow rate at outlet

52 © 2015 ANSYS, Inc. June 12, 2018

 Lee Model – Residuals (Varying Evap frequency)
1000

500
200
With increasing evaporation frequency,
convergence will be difficult to achieve

53 © 2015 ANSYS, Inc. June 12, 2018

 Lee Model – Varying Evaporation Frequency
Volume fraction @ outlet
500
200 1000
100

With varying evaporation frequencies,

4.0
vapor volume fraction at outlet changes
and hence, it has to be fine tuned to
match expected results

0.5
0.1

54 © 2015 ANSYS, Inc. June 12, 2018

 Solution - III
Best practices for Lee Evaporation Condensation Model
Search Keywords: evaporation
condensation model best practices
Customer Portal KM ID: 2028458

55 © 2015 ANSYS, Inc. June 12, 2018

 Evaporation Condensation(Lee) Model

• The Lee model is a mechanistic model with a physical basis.

• It is used with the mixture and VOF multiphase models and can be
selected with the Eulerian multiphase model if one of the overall
interfacial heat transfer coefficient models will be used (as opposed to the
two-resistance model)

56 © 2015 ANSYS, Inc. June 12, 2018

 Setting up Evaporation Condensation(Lee) Model
• For details on how to set up the mass transfer please refer to following
section in the documentation:
help/flu_ug/flu_ug_sec_mphase_massxfr.html

• Always specify the mass transfer from liquid to vapor

57 © 2015 ANSYS, Inc. June 12, 2018

 Evaporation Condensation(Lee) Model
Latent Heat
• Latent heat is taken care of through formation (standard state) enthalpy
specified in the materials panel.
• To correctly specify the latent heat, please specify following:
– For liquid and vapor, use reference temperature as saturation temperature.
– Specify the same molecular weight for liquid and vapor phase.
– For liquid specify standard state enthalpy = 0, and for vapor specify standard state enthapy =
latent heat (in j/kg-mol units)
– If you have latent heat data in j/kg then multiply it with molecular weight (unit kg/kg-mol). This
will give the latent heat value in required units.

• This specification will ensure appropriate latent heat removal/addition

• Standard state entropy has no role to play. It can be left to default
value.

58 © 2015 ANSYS, Inc. June 12, 2018

 Evaporation Condensation(Lee) Model
Calculation strategy
• If you are using mixture model for steady state calculation, you can go for coupled
solver with rather lower Courant numbers like 10 or even less at times.
• If you still face problems, lower the explicit relaxation factors for both pressure and
momentum to 0.75 or even up to 0.5.
• If coupled solver does not work at all, go for Pseudo transient.
• If steady calculation is causing issues, check if the reverse flow is causing any trouble. If
you are sure that there is no reverse flow in the final solution, specify reverse flow
quantities such that they help convergence. For example, if you initialize with all liquid,
then specify reversed flow of liquid. This will help convergence.
• If nothing works, go for unsteady calculation. For unsteady, use phase-coupled SIMPLE,
with following URF values. Pressure 0.5/0.6, momentum 0.3/0.4, vaporization mass 0.1
to 0.5, vof 0.5, density 0.5.

59 © 2015 ANSYS, Inc. June 12, 2018

 Evaporation Condensation(Lee) Model

• This is an approximate model in which departure from saturation determines the rate
of mass transfer. e.g. (Tcell - Tsat) is the driving force.
• This means, for mass transfer to take place, the cell temperature has to be a little
above or below saturation temperature.
• In the post processing, you will always find mixture temperature away from saturation
temperature. (above saturation for evaporation problems and below saturation for
condensation problems)
• Increasing the evaporation frequency would reduce the mixture temperature and bring
it closer to saturation for evaporation problems.
• Increasing the condensation frequency would increase the mixture temperature and
bring it closer to saturation for condensation problems.

60 © 2015 ANSYS, Inc. June 12, 2018

 Topic – Cavitation Modeling
• What is it?
− Creation of vapor bubbles in a liquid when the static pressure falls
below the saturation vapor pressure
• Often seen in rotating machinery (pumps), fuel injectors, and
other high pressure fluid devices.
• Can be very destructive due to bubble collapse on surfaces
• Cavitation also produces lower device performance, vibration,
and noise
− Physics
• Multiple phases, large density ratios, turbulence, mass transfer
• Multiphase species transport, non-condensable gases (e.g. air)
• Effects of slip velocities between the liquid and gas phases
• Thermal effects (property variations) and compressibility of both
liquid and gas phases

61 © 2015 ANSYS, Inc. June 12, 2018

 ANSYS Solutions

• Both CFX and Fluent have cavitation

modeling capability
− Tutorials, training notes, and documentation
resources available at Customer Portal
• Applications include
− Axial and centrifugal turbopumps
− Positive displacement pumps
− Fuel injectors
− Sharp-edged orifices

62 © 2015 ANSYS, Inc. June 12, 2018

 Cavitation Search – Tutorials & Training Materials

Resources for customers

who are new to modeling
cavitation. Links can be
provided in SRs and
customer e-mails.

63 © 2015 ANSYS, Inc. June 12, 2018

 Cavitation – Best Practices
Search Text: cavitation best practices

Customer Portal KM ID: 2034438

Question: Are there any best practices for cavitation modeling?

Solution: Fluent best practices for cavitation modeling with vapor as
compressible phase.

64 © 2015 ANSYS, Inc. June 12, 2018

 KM 2034438: Content

• PDF document
• Solver specific best practices and
recommended settings
• Good for experienced Fluent user
who is new to cavitation modeling
• Less experienced users would
benefit from training notes and
tutorials

65 © 2015 ANSYS, Inc. June 12, 2018

 Cavitation and Dynamic Meshes
Search Text: cavitation dynamic mesh

Customer Portal KM ID: 1040

Question: I am modeling a positive displacement pump. Is there any information on

cavitation modeling with dynamic meshes in Fluent?
Solution: Cavitation model with dynamic mesh doesn’t converge. Solution
provided in the KM.

66 © 2015 ANSYS, Inc. June 12, 2018

 KM 1040: Solution

This KM illustrates the type of

solution that is not available
in the online documentation

KM may be generally useful for more than just dynamic mesh models…

67 © 2015 ANSYS, Inc. June 12, 2018

 Cavitation and Non-condensable Gases
Search Text: cavitation non-condensable gases

Customer Portal KM ID: 2040009

Question: I am modeling a cavitating fuel injector, and would like to account for a
non-condensable gas (air) in the system. What is the approach for doing this in Fluent?
Solution: How to simulate noncondensable gases in conjunction with
cavitation.

68 © 2015 ANSYS, Inc. June 12, 2018

 KM 2040009: Solution

This KM illustrates a
recommendation based on
experience with these
problems at ANSYS

Three options:
1. Model air as a third phase.
2. Model vapor+air as a multi-species secondary phase
3. Activate the Singhal model, which includes effect of noncondensable gas

69 © 2015 ANSYS, Inc. June 12, 2018

 ANSYS Knowledge Database

Structures FAQs

Fluids Demos
17,000+ knowledge
Electromagnetics materials Tips & tricks

Installation/licensing Applications

Highlights Created, reviewed & maintained by ANSYS experts.

Over 500 video demos showcasing pre- and post-processing.

Used by over 40,000 people from around the world every month.

70 © 2015 ANSYS, Inc. June 12, 2018