Optional Lecture 3: Wall Boiling Models: 16.0 Release

Optional Lecture 3:
Wall Boiling Models

16.0 Release
Multiphase Modeling using

ANSYS Fluent
1 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential
Outline
• Introduction
• Theory
• Usage
• Validation Studies
• Best Practices
• Appendix

What is Boiling
• Boiling occurs in many industrial applications and
is characterized by large heat transfer coefficients
• This efficient heat transfer mechanism is limited

by the critical heat flux where the heat transfer
coefficient decreases leading to a rapid
temperature rise potentially leading to system
melting and destruction
• Traditionally, empirical correlations for critical

heat flux are used to develop an understanding of
the process dynamics
• But these correlations are valid only in the limited

specified region of fluid conditions and fluid
properties and hold for a defined geometry

Need of boiling simulation
Auto Industry
• Engine cooling
• Heat treatment
Phase change in micro channels
Power Industry
• Boiling in tubes or fuel rods
• Burn out simulation
Engine jacket cooling Electronics
• Electronic chip cooling
• Micro channel boiling
Electronic chip cooling
Quenching process
Burn out of electrically heated wire
Nuclear Power Plant

Boiling curve, applications & ANSYS capabilities
Auto
Power Critical Current ANSYS
Heat Flux Burnout Capabilities
Electronics
Available
Minimum
Heat Flux Unstable region –
Heat Flux
need no modeling
Subcooled Satu- Transitional Stable

rated or Unstable
Single Phase Nucleate boiling Film boiling
0 Wall Superheat (Twall - Tsat)

What happens at the wall?
CHF
Heat
Flux
MHF
Single
Phase Film
Nucleate boiling
boiling
Wall Superheat (Twall - Tsat)
liquid
Wall
Heat flux
Physical Model
• The conventional approach for modelling boiling process with significant volume fractions
of both phases is the Eulerian multiphase model
• Phase distributions are calculated by solving separate of continuity equations and

momentum equations for each phase
• For the steam-water flow an energy equation is solved for water, while vapour is
assumed to be saturated
– Optionally, solve energy equation for vapour as well
• The interactions between phases are modelled using the appropriate empirical models
for interfacial forces
– Drag, lift, turbulent dispersion and wall lubrication

RPI wall boiling algorithm and its extension
Singlephase flow with low q
Multiphase flow with high q Flow regime Treatment at the wall Treatment in the domain
Single Phase Heat flux to liquid No special

treatment
Multiphase Heat flux to liquid Mono Disperse flow

•subcooled •Euler model
q+
q++ +
qq++ nucleate boiling
Evaporation heat flux
Multiphase Heat flux to liquid Poly disperse flow

•saturated boiling, •Population balance
Evaporation heat flux
•slug-flow, •Interfacial area
•annular, Heat flux to vapor
concentration
•droplet flow
With ANSYS CFD, it is possible to model
• Subcooled nucleate boiling using RPI wall boiling model (CFX and Fluent)
• Saturated boiling using Non-equilibrium wall boiling model (Fluent)
• Critical Heat Flux (CHF)/Departure from Nucleate Boiling (DNB)/Burn Out conditions using CHF wall
Liquid boiling model (Fluent)

Subcooled Boiling
• Subcooled boiling occurs when the heat flux applied to the wall is too high to be transferred to
the bulk of liquid by the single phase convective and conductive processes
• The term “subcooled” means, that the saturation temperature is exceeded only in a local
vicinity of the wall, whereas the average temperature in the bulk remains below saturation
• Steam bubbles are generated at the heated surface at nucleation sites, with the surface
density of these sites depending various factors including the superheat ∆𝑻𝒔𝒂𝒕
• The attached bubbles grow and then leave the wall at certain critical size
– Depends on surface tension and flow regime of surrounding fluids
• Heat transfer from the wall is then described as being carried by turbulent convection of liquid,
by transient conduction due to the departing bubbles, and by evaporation

Subcooled Boiling: RPI Model
• The overall heat balance at the wall, the so called heat partitioning is usually written as
qW  qC  qQ  qE
 
Convective Quenching Evaporative
• The quenching heat flux represents the effects of transient conduction through the
patches of the fresh bulk liquid, coming to the wall to replace each departing steam
bubble
• Distribution of the entire wall heat flux between these mechanisms can be calculated by
modelling each mechanism in terms of:
– the nucleation site density,
– the size of departing bubble and their detachment frequency
– Waiting time for the next bubble to appear on the site

Subcooled Boiling: RPI Model – cont.
• The convective heat flux is correlated with the single phase heat transfer coefficient
multiplied by the area fraction (𝟏 − 𝑨𝒃 )
qC  hC ( TW  Tl )( 1  Ab )
• Similarly, the quenching heat flux is modelled as:

qQ  hQ Ab (TW  Tl )
2kl
hQ  l  Thermal Diffusivity, T  Periodic Time
lT
• The evaporation heat flux is obtained via the evaporation mass flux on the wall
qE  Vd N w ρv h fv f

Nucleate Boiling Model Parameters
• Area Fraction  Ja sat 
  NW DW2   
Ab  min 
1, K 
 
 , K  4.8e  80 
  4 
• Nucleation site density – this is correlated according to Lemmert and Chawla
NW  C n Tsat
m
n  210 m  1.805
– Other models : Kocamustafaogullari and Ishii
• Bubble departure diameter is correlated according to Tolubinsky and Kostanchuk
   Tsat  
DW  min
 0.0006  e 45 ,0.0014 

   
– Others models: Kocamustafaogullari and Ishii and Unal

• Bubble departure frequency
1 4 g
f  
T 3DW  p

Non-Equilibrium Model
• The RPI model does not calculate the temperature of the vapor, but instead it is fixed at
the saturation temperature
• To model boiling up to the critical heat flux and beyond, it is necessary to include the
vapour temperature in the solution process
• The wall heat partition is modified to allow for heat flux to vapour and other non-
condensable gases:
qwall  f 1 (qC  qQ  qE )  (1  f 1 )qV  qG
 
convective heat flux to vapour Other gas
• The function 𝒇𝜶𝟏 varies from 0 (Dry out situation) to 1 (subcooled nucleate boiling)

Critical Heat Flux Model
• When wall boiling departs from the nucleate boiling regime and reaches the critical heat flux
and post dry-out conditions, the multiphase flow regime changes from a bubbly flow to a mist
flow
• With the flow regime transition, the interfacial area, momentum transfer terms heat transfer
and turbulence quantities change
• The flow regimes are determined by a single local flow quantity - the vapour volume fraction
– If the 𝜶𝒑 =< 𝟎. 𝟑 → bubbly flow
– If the 𝜶𝒑 => 𝟎. 𝟕 → mist flow
– For all other (𝟎. 𝟑 > 𝜶𝒑 < 𝟎. 𝟕) → Churn flow
• The flow regimes are used to compute the interfacial area and interfacial transfers forces
Interfacial Forces
• The interfacial drag force is calculated using the standard drag model described
in Eulerian lecture. In addition for boiling flows the Ishii model is available and this is
typically chosen for modelling interphase exchange
• For boiling flows, turbulent dispersion force is important in transporting the vapour from
walls to the core fluid flow regions. In the RPI model the Lopez de Bertodano model is
usually chosen to account for the effects of the turbulent dispersion force
• Wall lubrication can be important in the nucleating boiling regime. In the RPI model, the
Antal et al. model is usually chosen to account for the effects of the wall lubrication force
• The heat transfer can be modelled using either the Ranz-Marshall or Tomiyama models
described

Available Wall Boiling Model
Wall Boiling Model Applicability
RPI Used for subcooled nucleate boiling. If the bulk
liquid temperature is below saturation temperature,
the flow is called subcooled flow. (Tsat – Tliquid > 3K)
Non-equilibrium Used for saturated boiling. If the bulk liquid
temperature is close to saturation temperature, the
flow is called saturated flow. (Tsat – Tliquid ≤ 3K)
Critical Heat Flux Used for Critical Heat Flux (CHF)/ Burn out/
Departure from Nucleate Boiling (DNB) situation

Boiling Sub-Models
Boiling Sub-Model Options
Bubble departure diameter tolubinski-konstanchuk, unal, kocamustafaogullari-ishii,
constant, user-defined
Frequency of Bubble cole, user-defined
Departure
Nucleation Site Density Lemmert-chawla, kocamustafaogullari-ishii, user-defined
Area Influence Coeff. delvalle-kenning, constant, user-defined
Quenching Model Fixed Yplus Value, Fixed Liquid Temperature
Correction
• Default sub-models are shown in bold. In general, they work well for variety of applications.
• Various sub-models include effect of different parameters, e.g. Unal correlation includes effect of operating
pressure on the bubble departure diameter. Please refer to the appendix/documentation for more details.
How to activate wall boiling models
• Wall boiling models are available with Eulerian multiphase
models panel
• There are three options available
– RPI Boiling Model
• Used for subcooled nucleate boiling. If the bulk liquid
temperature is below saturation temperature, the flow is
called subcooled flow. (Tsat – Tliquid > 3K)
– Non-equilibrium Boiling
• Used for saturated boiling. If the bulk liquid temperature is
close to saturation temperature, the flow is called saturated
flow. (Tsat – Tliquid ≤ 3K)
– Critical Heat Flux
• Used for Critical Heat Flux (CHF)/ Burn out/ Departure from
Nucleate Boiling (DNB) situation
• No special set up is required at the wall boundaries
– Please note that boiling will not work with a wall having fluids
on both the sides

How to set up a wall boiling model

How to set up a wall boiling model – cont.
Leave to default
User input
Select
appropriate
sub-models

Secondary phase diameter
• Please select liquid as primary and vapor
as secondary
• Please make sure to select boiling-dia as
the Diameter for secondary phase
• This diameter is a function of local sub-
cooling (Tsat - Tliq). More details in the
appendix.

Setting Tvapor = Tsat for RPI wall boiling
• For subcooled nucleate boiling,
– The vapor superheat is not very high (< 10K)
– The vapor volume fraction is also lower
• The heat capacity of vapor is lower than
water
• Due to above reasons, It is a common
practice to fix the vapor temperature to
saturation
• Fixed values option in the cell zone panel
can be used to fix the vapor temperature
to saturation

Subcooled Nucleate Boiling
Vertical pipe
Length: 2 m
Diameter: 15.4 mm
Heat Flux: 570 kW/m2
Mass Flux: 900 kg/m2-s
Operating pressure: 4.5 MPa
0.5
0.4
Experiments
0.3 ANSYS CFD (Fluent)
0.2 RPI_paper
0.1
0.0 Void fraction Temperature in K
0.0 0.5 1.0 1.5 2.0
Axial distribution of Average void fraction
• RPI wall boiling model used
• The RPI approach is applicable to subcooled nucleate boiling regime, particularly in vertical pipes/channels.
• This model is also be applied to other problems, as long as the regime is subcooled nucleate boiling.

Critical Heat Flux
CHF
Vertical pipe
Length: 7 m
Diameter: 10 mm

Best Practices
16.0 Release

Available material
Theory & Usage:
Refer to ANSYS R16 documentation:
• ANSYS Fluent 16.0 Theory Guide: Section 17.5.20. Wall Boiling Models
– Documentation link:- help/flu_th/flu_th_sec_boiling.html
• ANSYS Fluent 16.0 User’s Guide: 25.5.7. Including the Boiling Model
– Documentation link:- help/flu_ug/flu_ug_sec_use_bm.html
Tutorial:
• Available on customer portal
https://support.ansys.com/AnsysCustomerPortal/en_us/Knowledge%20Resources/Tutorial
s%20&%20Training%20Materials/Pdf%20Tutorial/Multiphase+Flow+Modeling+using+ANSY
S+FLUENT+Tutorials+:+Modeling+Nucleate+Boiling+Using+ANSYS+FLUENT

Best Practice on Set up
It is recommended to avoid very fine mesh near the boiling wall. Wall function meshes with
Y+> 30 work well with the boiling calculations. For boiling at low operating pressures (1
bar), first cell size of the order of ~1 mm works reasonably well.
Works with all variant of k-epsilon and k-omega. Boiling is not available with Laminar flows.
Common mistakes to avoid:
• Make sure that gravity is switched on when you solve boiling. Without gravity, calculation
may run but you may not see any heat transfer
• Make sure that in the phase interaction panel, surface tension is specified. For boiling
calculations, it is a must.
• In the phase interaction panel, when the Number of Mass Transfer Mechanisms is set to
1, From Phase and To Phase are set to primary phase by default. Make sure appropriate
phases are selected at this stage.

Best Practices on Numerics
• Boiling calculation can be run in steady state
• Pressure Velocity coupling as Coupled
• Select on the pseudo transient solver
– Lower the time scale factor if you still face convergence trouble. You can reduce it to as low as
0.001
• If steady calculation is causing issues, check if the reverse flow is causing any trouble. If
there is no reverse flow in the final solution, specify reverse flow quantities such that
they help convergence. For example, if the case is initialized with all liquid, then specify
reversed flow of liquid, although in final solution, it may be vapor. This will help
convergence.

Recommended Setting for Grid Independent Results
• With boiling calculations, using quenching correction
option is recommended
– Set the minimum reference temperature to the minimum
temperature expected in the domain (generally the inlet
temperature)
– Set the Yplus value to 250
• For fine meshes, using wall lubrication force may be
necessary. This option available in the phase
interaction panel
– Antal et. al. is recommended
– Refer to Appendix for more details

Appendix – Boiling Sub-Models
16.0 Release

Types of Boiling
• Subcooled boiling:
– Occurs when the heat flux applied to the wall is too high to be transferred to the bulk of liquid by the
single phase convective and conductive processes.
– The term “subcooled” means the average temperature in the bulk remains below saturation.
• Saturated Boiling
– Occurs when the bulk of the liquid reaches a boiling temperature.
– The boiling process at these conditions becomes violent
• Departure from Nucleate Boiling (DNB)
– Occurs when vapour volume fraction becomes large and forms a patchy layer covering the heated
surface
– Reduces the heat transfer to the core and it remains liquid
• Critical Heat Flux (CHF)/ Burn out
– Occurs when the core when assume a droplet regimes, liquid film dries out and vapour form continuous
layer on the heated surface.

Boiling sub-models – Area of Influence

Boiling sub-models – Area of Influence – cont.

Boiling sub-models – Frequency of bubble
departure

Boiling sub-models – Nucleate Site Density

Boiling sub-models – Bubble departure
diameter

diameter – cont.

diameter – cont.

Boiling sub-models – Non-equilibrium
Subcooled Boiling

Boiling sub-models – Critical Heat Flux

Boiling sub-models – Critical Heat Flux – cont.

Boiling sub-models – Flow Regime Transition

Boiling sub-models – Interfacial Transfer
• For interfacial heat and momentum transfer, it is required to calculate the dispersed
phase diameter
• For bubbly flow the bubble diameter and the mist flow the droplet diameter are required

Boiling sub-models – bubble diameter

Boiling sub-models – Droplet diameter

Boiling sub-models – interfacial heat transfer

Boiling sub-models – Interfacial Mass Transfer

Appendix – Ensuring Grid
Independence for Boiling Simulation
16.0 Release

default settings for boiling model

Results with Default settings
Problem: Vertical pipe
Length: 2 m
0.6 Diameter: 15.4 mm
0.5
0.4
Experiments
0.3 RPI_paper
_mesh1
F14
mesh2
F14_adapted
0.2 mesh3
F14_2level_adapted
0.1
0
0.00 0.40 0.80 1.20 1.60 2.00
Observations & Reasons for Mesh Dependency
Observations
• Mesh dependency
• Excessive vapor generation close to “Onset of boiling”
Reasons:
q wall  q f  qe  qq
• qf = Single phase heat transfer to liquid

– Grid independent as it uses a heat transfer coefficient for liquid calculated by the solver internally from
wall functions. This has in-built grid independence treatment.
• qe = Evaporation heat flux, f(Twall – Tsat)
– Grid independent as it does not use liquid temperature from the cell next to the wall
• qq = Quenching heat flux, f(Twall – Tliq)
– Grid dependent component, as it requires Tliq taken from the near wall cell. In original correlation, it is
taken from the centerline temperature, however, this is not possible in a 3D CFD calculation.

Solution used in CFX
• qq = Quenching heat flux, f(Twall – Tliq)
• In CFX, Following is implemented to make this component grid independent
• Based on the temperature in the cell next to the wall and its Y+, liquid temperature at
Y+=250 is estimated and used in the above correlation.
• qq = Quenching heat flux, f(Twall – Tliq_at_Y+=250)
• This option is available in Fluent14 as Quenching correction.
– There are two options for Tliq.
• Liquid temperature at Y+=250 can be used. This is the preferred option.
• Alternatively, a constant temperature can be specified

Quenching correction options in Fluent

Checking grid independence for quenching
correction option Y+=250
0.6
0.5
0.4
Experiments
RPI_paper
0.3
F14,_mesh1
fixed Y+ = 250
mesh2 fixed Y+=250
F14_adapted,
0.2 _mesh3
F14 2 level adapted, fixed Y+=250
Still some deviation seen,

0.1 especially with fine meshes.
This is because modeling the
wall lubrication force
0 becomes more important
0.00 0.40 0.80 1.20 1.60 2.00 with fine meshes.
Results with Wall lubrication
0.6
0.5
0.4
Experiments
RPI_paper
0.3
F14,_mesh1
fixed Y+ = 250
mesh2
F14_adapted,fixed Y+=250,WL
0.2
_mesh3
F14 2 level adapted, fixed Y+=250, WL
0.1
0
0.00 0.40 0.80 1.20 1.60 2.00

Summarizing Quenching Correction
• With boiling calculations, using quenching correction with Y+ = 250 option is
recommended
• For fine meshes, using wall lubrication force may be necessary. This option available in
the phase interaction panel

Optional Lecture 3: Wall Boiling Models: 16.0 Release

Uploaded by

Copyright:

Available Formats

Optional Lecture 3: Wall Boiling Models: 16.0 Release

Uploaded by

Document Information

Original Title

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

Optional Lecture 3: Wall Boiling Models: 16.0 Release

Uploaded by

Copyright:

Available Formats

Optional Lecture 3:

Wall Boiling Models

Multiphase Modeling using

2 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

• This efﬁcient heat transfer mechanism is limited

• Traditionally, empirical correlations for critical

• But these correlations are valid only in the limited

3 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

4 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

Subcooled Satu- Transitional Stable

Single Phase Nucleate boiling Film boiling

0 Wall Superheat (Twall - Tsat)

• Phase distributions are calculated by solving separate of continuity equations and

7 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

Single Phase Heat flux to liquid No special

Multiphase Heat flux to liquid Mono Disperse flow

Multiphase Heat flux to liquid Poly disperse flow

8 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

11 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

12 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

• Similarly, the quenching heat flux is modelled as:

13 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

– Others models: Kocamustafaogullari and Ishii and Unal

14 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

15 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

17 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

18 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

20 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

21 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

22 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

23 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

24 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

25 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

26 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

27 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

28 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

29 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

30 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

31 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

32 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

33 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

34 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

35 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

36 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

37 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

38 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

39 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

40 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

41 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

42 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

43 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

44 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

45 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

46 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

47 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

48 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

49 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

50 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential

51 © 2015 ANSYS, Inc. December 15, 2015 ANSYS Confidential