Best Practices for

Modeling Heat
Transfer in Simcenter
STAR-CCM+
Steven Dowding

Unrestricted © Siemens AG 2018

Overview

• Heat Transfer in Simcenter STAR-CCM+

• Geometry Preparation
• Meshing
• Parts Based Workflow
• Convergence
• Heat Transfer Coefficients
• Summary

Unrestricted © Siemens AG 2018

Page 2 2018-10-17 Siemens PLM Software
Overview

• Heat Transfer in Simcenter STAR-CCM+

• Geometry Preparation
• Meshing
• Parts Based Workflow
• Convergence
• Heat Transfer Coefficients
• Summary

Unrestricted © Siemens AG 2018

Page 3 2018-10-17 Siemens PLM Software
Heat Transfer in Simcenter STAR-CCM+ - Why Is Heat
Transfer Important?

• Improved efficiency
Improved performance • Emissions reduction
• Smaller packaging

• Wide operability
Increased reliability • Reduce thermal stress
• Longer component lifespan

• Reduce expensive field testing

Cost management • Improved operational efficiencies
• Reduced maintenance costs

Unrestricted © Siemens AG 2018

Page 4 2018-10-17 Siemens PLM Software
Heat Transfer in Simcenter STAR-CCM+

• In the context of Simcenter STAR-CCM+, when discussing

heat transfer, we are generally referring to Conjugate Heat
Transfer (CHT)
• Multiple modes of heat transfer present in a system:
• Conduction
• Convection
• Radiation

• What does this mean for us in terms of simulation setup?

• Multiple Material Properties
• Solids, liquids, gases may all be present
• Multiple Parts/Regions
• Communication must be allowed between them
• Multiple Timescales
• Solid temperature changes slow compared to fluid

Unrestricted © Siemens AG 2018

Page 5 2018-10-17 Siemens PLM Software
Heat Transfer in Simcenter STAR-CCM+ - Solver
Considerations

• We must turn our raw geometry into a discrete representation in

order to solve our numerical equations
• This is the process of meshing
• How this mesh is generated is important
• Balance between numerical accuracy/solver requirements and
engineer time needed to construct the mesh
• Two main types of meshes can be generated, each with its own
implications for how information is exchanged
• Conformal
• One-to-one node matching, best accuracy from solver
• May require more time to prepare geometry to meet
requirements
• Non-Conformal Mesh
• Mismatch of nodes, solver needs to perform some type of
face matching or interpolation
• Less stringent geometry requirements to generate
Unrestricted © Siemens AG 2018
Page 6 2018-10-17 Siemens PLM Software
Overview

• Heat Transfer in Simcenter STAR-CCM+

• Geometry Preparation
• Meshing
• Parts Based Workflow
• Convergence
• Heat Transfer Coefficients
• Summary

Unrestricted © Siemens AG 2018

Page 7 2018-10-17 Siemens PLM Software
Geometry Preparation – Enabling Communication Between
Parts

• With multiple materials present, we need to enable communication between the

various parts that represent them
• In STAR-CCM+, we create Interfaces between parts to exchange information

Solid F
l
Interface u
x
Fluid
• We have several different options for how we create these interfaces:
1. Conformal Imprint
2. Non-Conformal Imprint
3. Weak Contact

Unrestricted © Siemens AG 2018

Page 8 2018-10-17 Siemens PLM Software
Geometry Preparation – CAD Imprinting - Conformal

• When working with clean, analytical CAD, use a CAD based imprint
• Higher quality due to use of underlying CAD representation
• Not affected by tessellation density
• Does not modify geometry topology
• Will not move surfaces to make them coincident
• Allows for use of Directed Mesher due to retention of CAD
• Generates a conformal mesh

Unrestricted © Siemens AG 2018

Page 9 2018-10-17 Siemens PLM Software
Geometry Preparation – Discrete Imprinting –
Conformal/Non-Conformal

• Use Discrete Imprint when not working with CAD

or coincident surfaces
• Able to move surfaces to make them coincident
at the interface
• Cannot be used with Directed Mesher (use
Extruder instead)
• Dependent on input
tessellation/tolerance
• Higher tessellation
density can increase
imprint quality
• May be difficult if
varying tolerances are
present
• Imprint can be conformal
Poor Imprint Good Imprint
or non-conformal
Unrestricted © Siemens AG 2018
Page 10 2018-10-17 Siemens PLM Software
Geometry Preparation – Imprinting Tip – Splitting Surfaces

• Imprinting can be used to split surfaces as well

• Automatically generate source and target
surfaces for Directed Mesher
• Automatically separate interfaces into separate
surfaces for post processing or custom mesh
controls
• Set Resulting Part Surfaces to New Part Surfaces

Unrestricted © Siemens AG 2018

Page 11 2018-10-17 Siemens PLM Software
Geometry Preparation – Imprinting Tip – Splitting Surfaces

Original Part Surfaces Resulting Part Surfaces

Unrestricted © Siemens AG 2018

Page 12 2018-10-17 Siemens PLM Software
Geometry Preparation – Weak Contacts

• The Weak Contact Creator can be used to generate interfaces between parts
without imprinting them
• This means that surfaces are not moved
• Can be more robust (less “breaking” of complex geometries)
• Surfaces are also not split
• No irregular feature curves which can be generated by an imprint
• Interfaces are always non-conformal in nature
• Essentially tells STAR-CCM+ that the two parts are touching each other, but
not exactly where
• Lets the solver decide later exactly where the contact occurs

Unrestricted © Siemens AG 2018

Page 13 2018-10-17 Siemens PLM Software
Geometry Preparation Overview

Conformal Non-Conformal Weak Contact

• Can use CAD based imprint if • Use non-conformal Discrete • Will not move or split surfaces
CAD is available for higher Imprint • Essentially allows the solver to
quality • Will not move surfaces to compute the exact interface
• Can use Discrete Imprint for make them coincident, but will later on
tessellated surfaces or when still split surfaces • Best used with complex
small tolerances are present • Best used if tolerances are geometries or those with
• Will move the surfaces to small, but moving surfaces larger/varying tolerances
make them coincident would cause errors in • Use in conjunction with
geometry Mapped Contact Interface

Unrestricted © Siemens AG 2018

Page 14 2018-10-17 Siemens PLM Software
Overview

• Heat Transfer in Simcenter STAR-CCM+

• Geometry Preparation

• Meshing
• Parts Based Workflow
• Convergence
• Heat Transfer Coefficients
• Summary

Unrestricted © Siemens AG 2018

Page 15 2018-10-17 Siemens PLM Software
Meshing

• When deciding how to generate our mesh, we need to consider

• How did we prepare the geometry
• What type of mesh/interfaces are we trying to achieve
• As with geometry preparation, we have three main paths to consider
1. Conformal mesh/interfaces
2. Non-conformal mesh/interfaces
3. Non-conformal mesh/Mapped Contact Interface (weak contact)
• Which one of these options is most appropriate depends on how we have
prepared our geometry, and what our domain looks like

Unrestricted © Siemens AG 2018

Page 16 2018-10-17 Siemens PLM Software
Meshing – Conformal Meshing/Interfaces

A conformal mesh is desirable for accuracy, so what

exactly is a conformal mesh?
• Cells share the same vertices at the interface
• One-to-one face match
• Smooth heat transfer across regions
• Not much to worry about in terms of the interfaces

Initialize

Unrestricted © Siemens AG 2018

Page 17 2018-10-17 Siemens PLM Software
Meshing – Conformal Meshing

So, a conformal mesh is preferred…..how do I

generate one?

Requirements:
• Geometry must be imprinted
• Either CAD imprint or Discrete conformal
• Parts must be in the same mesh operation
• Per-part/concurrent meshing disabled
• Polyhedral mesher must be used
• Trimmed cell mesher does not guarantee
conformality between regions

Unrestricted © Siemens AG 2018

Page 18 2018-10-17 Siemens PLM Software
Meshing – Non-conformal Meshing

• What if I can’t generate a conformal mesh?

• Sometimes, geometry is too complex or dirty to
create conformal interfaces
• Per-part or concurrent meshing is desired for
efficient meshing or large numbers of parts
• That’s OK!
• Use non-conformal imprint
• STAR-CCM+ will automatically attempt to split
faces at the interface in order to achieve a match

Initialize

Unrestricted © Siemens AG 2018

Page 19 2018-10-17 Siemens PLM Software
Meshing – Non-Conformal Meshing

• When generating a non-conformal mesh, face

Unmatched faces remain as
matching at the interface is important
walls – no heat transfer!
• If there are large gaps or changes in mesh
resolution across the interface, the solver may
not be able to split and match all faces
• Umatched faces will not be included in interface
calculations
• Effectively treated as walls, meaning they are
adiabatic
• These faces can be visualized in a mesh scene
• Unmatched faces show as default grey wall

Unrestricted © Siemens AG 2018

Page 20 2018-10-17 Siemens PLM Software
Meshing – Non-Conformal Meshing

• Ensure high matching across the interface (>98%)

• Generating similar cell sizes on both sides of the
interface can help in achieving higher matching
• Can increase interface intersection tolerance to
attain greater matching
• Generally, if you need to go higher than 0.1,
should consider repairing the geometry or using
a Weak Contact/Mapped Contact Interface
instead

Unrestricted © Siemens AG 2018

Page 21 2018-10-17 Siemens PLM Software
Meshing – Non-Conformal Weak Contact

• Useful for very dirty geometries (where moving or splitting faces

would cause issues) and when good matching using a regular non-
conformal interface cannot be achieved
• Can make use of concurrent meshing
• Attempts no interface intersection
• Uses proximity based data mapper to calculate heat transfer
across interface (Mapped Contact Interface)
• Not guaranteed to conserve energy

Initialize

Unrestricted © Siemens AG 2018

Page 22 2018-10-17 Siemens PLM Software
Meshing – Non-Conformal Weak Contact

• Similar to a regular non-conformal mesh, ensuring similar cell sizes on both

sides of the interface is important
• Helps ensure more accurate mapping and energy conservation

Unrestricted © Siemens AG 2018

Page 23 2018-10-17 Siemens PLM Software
Meshing – Mesh Model Setup

• Polyhedral Mesher Recommended

• More effective for complex flows
• Accurately captures complex geometric shapes
• Can be used in conjuction with Thin Mesher
• Can generate a conformal mesh

Unrestricted © Siemens AG 2018

Page 24 2018-10-17 Siemens PLM Software
Meshing – Thin Meshing

• Thin Mesher should be used when thin parts are

present
• Produces higher quality prismatic cells which fit
better into thin areas
• Reduces mesh count compared to packing a
large number of very small poly cells to produce
the desired mesh resolution

Unrestricted © Siemens AG 2018

Page 25 2018-10-17 Siemens PLM Software
Meshing – Thin Meshing

• Thin Mesher automatically

detects thin regions, and
generates prismatic cells
• Automatically transitions back to
poly in bulk areas
• Mesh is fully conformal (in
transition from thin to poly cells)
as well as between parts (if all
other requirements are met)

Unrestricted © Siemens AG 2018

Page 26 2018-10-17 Siemens PLM Software
 Meshing – Mesh Quality

• Ensure smooth transition from prisms to core volume mesh – enough layers to resolve boundary layer
• Setting Near Core Layer Aspect Ratio (NCLAR) on Prism Layer mesher can help with this (0.5-1.0)
• Will reduce prism layer thickness to prevent last prism layer from having higher aspect ratio than first
volume cell

BAD BETTER BEST

NCLAR= 0, NCLAR= 1.0, NCLAR= 1.0, Added Layers
Bad Transition, Bad Resolution Better Transition, Bad Resolution Better Transition, Better Resolution
Unrestricted © Siemens AG 2018
Page 27 2018-10-17 Siemens PLM Software
Meshing Overview

Conformal Non-Conformal Weak Contact/Mapped Interface

• Highest geometry preparation • Fewer requirements on • Least requirements on
requirements geometry preparation (do not geometry preparation
• Most accurate from solver need to make surfaces • Will not move or split surfaces
standpoint coincident) • Essentially allows the solver to
• Requires conformal imprint • Requires non-conformal compute the exact interface at
(either CAD or Discrete) Discrete Imprint runtime
• All parts must be meshed in • Can make use of concurrent • Best used with complex
the same operation using meshing geometries or those with
Polyhedral Mesh • Can make use of other mesh larger/varying tolerances
• Cannot make use of types, such as Trimmed Cell • Most due diligence required at
concurrent meshing for large Mesher the interface to ensure energy
assemblies (will break • More due diligence needed to conservation due to use of
conformality) ensure high face matching on mapping approach
interface

Unrestricted © Siemens AG 2018

Page 28 2018-10-17 Siemens PLM Software
Overview

• Heat Transfer in Simcenter STAR-CCM+

• Geometry Preparation
• Meshing

• Parts Based Workflow

• Convergence
• Heat Transfer Coefficients
• Summary

Unrestricted © Siemens AG 2018

Page 29 2018-10-17 Siemens PLM Software
Parts Based Workflow

• Why parts based?

• Improves scalability and usability, both in terms of hardware usage and
engineer interaction
• Clearer simulation setup
• Less clutter
• Easier sim tree navigation
• Less entities requiring engineer’s time to setup
• Improved scalability
• Numerically more efficient for the solver
• Improvements realized through several reductions
• Reduced number of Continua, Regions, and Boundaries
• Single region can represent multiple solid materials
• Single continua can represent multiple solid physics
• Single interface can represent multiple part contacts
• Less engineer time required to perform setup and provide inputs due to
simplification of the user interface
Unrestricted © Siemens AG 2018
Page 30 2018-10-17 Siemens PLM Software
 Parts Based Workflow – Motivation: VTM Case

Reduced solution time for large assemblies

• Allows consolidation of regions, boundaries and interfaces
• Benefits seen in electronics cooling, vehicle
thermal management, steam generators etc.

Boundary Based Interfaces VTM

Parts Based Interfaces
60 53.9M cells
Interfaces 8600 > 28
50 Boundaries 37102 > 80
ELAPSED TIME [S]

Regions 5048 > 1

40 Continua 167 > 1

30

20

10

0
8 16 32 64
Unrestricted © Siemens AG 2018
CORES
Page 31 2018-10-17 Siemens PLM Software
Parts Based Setup – Generic Heat Exchanger

Hot Side Inlet Hot Side Outlet

Inlet Hose Outlet Hose

Cold Side Flow Cooling Tubes

Inlet Tank Outlet Tank

Unrestricted © Siemens AG 2018

Page 32 2018-10-17 Siemens PLM Software
Parts Based Setup – Generic Heat Exchanger Boundary
Conditions

Mass Flow Inlet Velocity Inlet and Pressure Outlet Pressure Outlet

Unrestricted © Siemens AG 2018

Page 33 2018-10-17 Siemens PLM Software
Parts Based Setup – Generic Heat Exchanger

Rubber
Air Copper

Plastic

Unrestricted © Siemens AG 2018

Page 34 2018-10-17 Siemens PLM Software
Parts Based Setup – Generic Heat Exchanger Materials
Solid Physics Continua

Multi-Part Solid Continuum

Unrestricted © Siemens AG 2018

Page 35 2018-10-17 Siemens PLM Software
Parts Based Setup – Generic Heat Exchanger Regions

Unrestricted © Siemens AG 2018

Page 36 2018-10-17 Siemens PLM Software
Parts Based Setup – Generic Heat Exchanger Interfaces

Unrestricted © Siemens AG 2018

Page 37 2018-10-17 Siemens PLM Software
Parts Based Setup - Generic Heat Exchanger Results

Region Parts
Based Based
Continua 4 2
Regions 55 2
Boundaries 599 13
Interfaces 182 3
Cells 3872543 3872543

Unrestricted © Siemens AG 2018

Page 38 2018-10-17 Siemens PLM Software
Parts Based Setup – Generic Heat Exchanger Performance

Region Based vs Parts Based Speedup

Region Based Setup Parts Based Setup Ideal
256

224

192
~168%
160
Speedup

128

96

64

32
~27% ~3.8 Million Cells
0
0 32 64 96 128 160 192 224 256
Number of Cores

Unrestricted © Siemens AG 2018

Page 39 2018-10-17 Siemens PLM Software
Overview

• Heat Transfer in Simcenter STAR-CCM+

• Geometry Preparation
• Meshing
• Parts Based Workflow

• Convergence
• Heat Transfer Coefficients
• Summary

Unrestricted © Siemens AG 2018

Page 40 2018-10-17 Siemens PLM Software
Convergence - Under Relaxation Factors

• Under Relaxation Factors (URF’s) effect the speed and stability of

convergence
• Too high – solution may become unstable, or diverge
• Too low – solution takes too long to converge

Unrestricted © Siemens AG 2018

Page 41 2018-10-17 Siemens PLM Software
Convergence - Under Relaxation Factors

• Solid Energy URF can be increased

• Default value of 0.99 is stable, but causes slow
convergence
• Can be increased to 0.9999 or 1.0

Unrestricted © Siemens AG 2018

Page 42 2018-10-17 Siemens PLM Software
Convergence – Effect of Solid Energy URF

Solid Energy URF = 0.99 Solid Energy URF = 1.0

Unrestricted © Siemens AG 2018

Page 43 2018-10-17 Siemens PLM Software
Convergence – Monitoring Convergence

• How do we know solution has converged?

• Are residuals enough?

• Lets take a look at our previous example

Unrestricted © Siemens AG 2018

Page 44 2018-10-17 Siemens PLM Software
Convergence – Monitoring Convergence

• Looking at the residuals, we might think this

simulation is converged
• Values are low
• Most have dropped a few orders of magnitude

• But lets take a deeper look at some engineering

quantities of interest

Unrestricted © Siemens AG 2018

Page 45 2018-10-17 Siemens PLM Software
Convergence – Monitoring Convergence

• Maximum Temperatures
• Use Maximum Report
• Monitor near locations of interest
• Should asymptote to a physically reasonable
value

Not Changing
Conclusion: Converged

Unrestricted © Siemens AG 2018

Page 46 2018-10-17 Siemens PLM Software
Convergence – Monitoring Convergence

• Minimum Temperatures
• Use Minimum Report
• Monitor near locations of interest
• Should asymptote to a physically reasonable
value

Still Changing
Conclusion: Not Converged

Unrestricted © Siemens AG 2018

Page 47 2018-10-17 Siemens PLM Software
Convergence – Monitoring Convergence

• Heat Transfer
• Use Heat Transfer Report on all
regions/boundaries
• Should asymptote to 0 if no energy sources
present
• Should asymptote to value of energy source if
source is present

Still Changing, not 0

Conclusion: Not Converged

Unrestricted © Siemens AG 2018

Page 48 2018-10-17 Siemens PLM Software
Convergence – Monitoring Convergence Overview

• Don’t just look at residuals!

• Monitor physical engineering quantities
• Values should asymptote to meaningful values
• Should monitor values from both the flow and energy fields
• Energy may converge at a different rate than the flow
• Similar method can be used for transient cases
• Values within a time step should asymptote

• Note the Sign Convention for heat transfer reports/field functions

• Heat going out of a boundary/interface = positive (based on outward facing
surface normals)

Unrestricted © Siemens AG 2018

Page 49 2018-10-17 Siemens PLM Software
Overview

• Heat Transfer in Simcenter STAR-CCM+

• Geometry Preparation
• Meshing
• Parts Based Workflow
• Convergence

• Heat Transfer Coefficients

• Summary

Unrestricted © Siemens AG 2018

Page 50 2018-10-17 Siemens PLM Software
Heat Transfer Coefficients

Heat Transfer Coefficients are calculated using the general form:

𝑞′′𝑠 = ℎ(𝑇𝑠𝑢𝑟𝑓𝑎𝑐𝑒 − 𝑇𝑟𝑒𝑓 )
Several different variations are available for post processing Heat Transfer
Coefficients
• Local Heat Transfer Coefficient
• Uses the near wall cell temperature as Tref to compute ΔT
• Heat Transfer Coefficient
• User Specifies Tref explicitly
• Specified y+ Heat Transfer Coefficient
• User Specifies y+, fluid properties at this y+ value are used to calculate h,
and determine Tref
• Virtual Local Heat Transfer Coefficient
• Does not need to solve the energy equation
• User specifies Tref and fluid properties

Unrestricted © Siemens AG 2018

Page 51 2018-10-17 Siemens PLM Software
Heat Transfer Coefficients

Local HTC Specified y+ HTC

Heat Transfer Coefficient Virtual Local HTC

Unrestricted © Siemens AG 2018
Page 52 2018-10-17 Siemens PLM Software
Heat Transfer Coefficients

• So which Heat Transfer Coefficient should I use?

• Choice of HTC is a post processing choice only and does not affect the
solution results computed internally when solving CHT in Simcenter STAR-
CCM+
• Should match HTC definition to what was used for comparison

• When coupling to other codes to solve for temperatures, generally Specified

y+ HTC is recommended
• Less dependent on mesh than Local HTC
• Gives better conditioned system at solve time
• Generally recommended to use y+ value of 100

Unrestricted © Siemens AG 2018

Page 53 2018-10-17 Siemens PLM Software
Overview

• Heat Transfer in Simcenter STAR-CCM+

• Geometry Preparation
• Meshing
• Parts Based Workflow
• Convergence
• Heat Transfer Coefficients

• Summary

Unrestricted © Siemens AG 2018

Page 54 2018-10-17 Siemens PLM Software
Summary

• Three main workflows for geometry prep/meshing depending on model needs

• Conformal/Imprinted
• Non-Conformal/Imprinted
• Non-Conformal/Weak Contacts
• Parts Based Workflows
• Improve scalability
• Reduce complexity in simulation setup
• URF’s can be modified to help speed convergence
• Set Solid Energy URF to 0.9999 or 1.0
• Engineering quantities of interest should be monitored for convergence
• Min/Max temperatures
• Energy Balance via Heat Transfer Report
• Choice of HTC is a purely post-processing choice
• When coupling to external codes, use Specified y+ HTC

Unrestricted © Siemens AG 2018

Page 55 2018-10-17 Siemens PLM Software
Thank you
Relevant Knowledge Base Articles

Heat Transfer Coefficients

Wall Functions in Heat Transfer Applications
Natural Convection Best Practices
Surface to Surface Radiation Best Practices
Tips for Meshing Thin Solids
Best Practices for Full Vehicle Thermal Setup
Best Practices for Modeling Heat Exchangers
Multi-Part Solid and Parts Based Profiles
Benefits of Contact Mode Interfaces
Imprinting Tips and Tricks

Unrestricted © Siemens AG 2018

Page 57 2018-10-17 Siemens PLM Software