Fluent Problems

FLUENT 6.
2 Tutorial Guide
January 2005
Copyright c 2005 by Fluent Inc.
All rights reserved. No part of this document may be reproduced or otherwise used in
any form without express written permission from Fluent Inc.
Airpak, FIDAP, FLUENT, FloWizard, GAMBIT, Icemax, Icepak, Icepro, MixSim, and
POLYFLOW are registered trademarks of Fluent Inc. All other products or name
brands are trademarks of their respective holders.
CHEMKIN is a registered trademark of Reaction Design Inc.
Portions of this program include material copyrighted by PathScale Corporation

2003-2004.
Fluent Inc.
Centerra Resource Park
10 Cavendish Court
Lebanon, NH 03766
Volume 1
1 Introduction to Using FLUENT: Fluid Flow and Heat Transfer in a Mixing Elbow
2 Modeling Periodic Flow and Heat Transfer
3 Modeling External Compressible Flow
4 Modeling Unsteady Compressible Flow
5 Modeling Radiation and Natural Convection
6 Using a Non-Conformal Mesh
7 Using a Single Rotating Reference Frame
8 Using Multiple Rotating Reference Frames
9 Using the Mixing Plane Model
10 Using Sliding Meshes
11 Using Dynamic Meshes
Volume 2
12 Modeling Species Transport and Gaseous Combustion
13 Using the Non-Premixed Combustion Model
14 Modeling Surface Chemistry
15 Modeling Evaporating Liquid Spray
16 Using the VOF Model
17 Modeling Cavitation
18 Using the Mixture and Eulerian Multiphase Models
19 Using the Eulerian Multiphase Model for Granular Flow
20 Modeling Solidification
21 Using the Eulerian Granular Multiphase Model with Heat Transfer
22 Postprocessing
23 Turbo Postprocessing
24 Parallel Processing
Using This Manual
What’s In This Manual

The FLUENT Tutorial Guide contains a number of tutorials that teach you how to use
FLUENT to solve different types of problems. In each tutorial, features related to problem
setup and postprocessing are demonstrated.
Tutorial 1 is a detailed tutorial designed to introduce the beginner to FLUENT. This
tutorial provides explicit instructions for all steps in the problem setup, solution, and
postprocessing. The remaining tutorials assume that you have read or solved Tutorial 1,
or that you are already familiar with FLUENT and its interface. In these tutorials, some
steps will not be shown explicitly.
All of the tutorials include some postprocessing instructions, but Tutorial 22 is devoted
entirely to standard postprocessing, and Tutorial 23 is devoted to turbomachinery-specific
postprocessing.
Where to Find the Files Used in the Tutorials

Each of the tutorials uses an existing mesh file. (Tutorials for mesh generation are
provided with the mesh generator documentation.) You will find the appropriate mesh
file (and any other relevant files used in the tutorial) on the FLUENT documentation CD.
The “Preparation” step of each tutorial will tell you where to find the necessary files.
(Note that Tutorials 22, 23, and 24 use existing case and data files.)
Some of the more complex tutorials may require a significant amount of computational
time. If you want to look at the results immediately, without waiting for the calcula-
tion to finish, you can find the case and data files associated with the tutorial on the
documentation CD (in the same directory where you found the mesh file).
How To Use This Manual

Depending on your familiarity with computational fluid dynamics and Fluent Inc. soft-
ware, you can use this tutorial guide in a variety of ways.
For the Beginner

If you are a beginning user of FLUENT you should first read and solve Tutorial 1, in order
to familiarize yourself with the interface and with basic setup and solution procedures.

c Fluent Inc. January 11, 2005 i
Using This Manual
You may then want to try a tutorial that demonstrates features that you are going to
use in your application. For example, if you are planning to solve a problem using the
non-premixed combustion model, you should look at Tutorial 13.
You may want to refer to other tutorials for instructions on using specific features, such
as custom field functions, grid scaling, and so on, even if the problem solved in the
tutorial is not of particular interest to you. To learn about postprocessing, you can look
at Tutorial 22, which is devoted entirely to postprocessing (although the other tutorials
all contain some postprocessing as well). For turbomachinery-specific postprocessing, see
Tutorial 23.
For the Experienced User

If you are an experienced FLUENT user, you can read and/or solve the tutorial(s) that
demonstrate features that you are going to use in your application. For example, if you
are planning to solve a problem using the non-premixed combustion model, you should
look at Tutorial 13.
You may want to refer to other tutorials for instructions on using specific features, such
as custom field functions, grid scaling, and so on, even if the problem solved in the
tutorial is not of particular interest to you. To learn about postprocessing, you can look
at Tutorial 22, which is devoted entirely to postprocessing (although the other tutorials
all contain some postprocessing as well). For turbomachinery-specific postprocessing, see
Tutorial 23.
Typographical Conventions Used In This Manual

Several typographical conventions are used in the text of the tutorials to facilitate your
learning process.
• An informational icon ( i ) marks an important note.
• An warning icon ( ! ) marks a warning.

• Different type styles are used to indicate graphical user interface menu items and
text interface menu items (e.g., Zone Surface panel, surface/zone-surface com-
mand).
• The text interface type style is also used when illustrating exactly what appears on
the screen or exactly what you must type in the text window or in a panel.
• Instructions for performing each step in a tutorial will appear in standard type.
Additional information about a step in a tutorial appears in italicized type.
ii
c Fluent Inc. January 11, 2005
Using This Manual
• A mini flow chart is used to indicate the menu selections that lead you to a specific
command or panel. For example,
Define −→Boundary Conditions...
indicates that the Boundary Conditions... menu item can be selected from the Define
pull-down menu.
The words surrounded by boxes invoke menus (or submenus) and the arrows point
from a specific menu toward the item you should select from that menu.

c Fluent Inc. January 11, 2005 iii
Using This Manual
iv
Contents
1 Introduction to Using FLUENT: Fluid Flow and Heat Transfer

in a Mixing Elbow 1-1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-1
Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-1
Problem Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-2
Setup and Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-3
Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-3
Step 1: Grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-4
Step 2: Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-9
Step 3: Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-11
Step 4: Boundary Conditions . . . . . . . . . . . . . . . . . . . . . . . . 1-13
Step 5: Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-18
Step 6: Displaying the Preliminary Solution . . . . . . . . . . . . . . . . 1-25
Step 7: Enabling Second-Order Discretization . . . . . . . . . . . . . . . 1-38
Step 8: Adapting the Grid . . . . . . . . . . . . . . . . . . . . . . . . . . 1-43
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-52
2 Modeling Periodic Flow and Heat Transfer 2-1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-1
Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-1
Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-3
Step 1: Grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-3
Step 2: Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-6

c Fluent Inc. January 11, 2005 i
CONTENTS
Step 3: Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-8

Step 5: Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-13
Step 6: Postprocessing . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-16
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-26
3 Modeling External Compressible Flow 3-1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-1
Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-1
Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-2
Step 1: Grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-3
Step 2: Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-6
Step 3: Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-8
Step 4: Operating Conditions . . . . . . . . . . . . . . . . . . . . . . . . 3-10
Step 6: Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-12
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-29
4 Modeling Unsteady Compressible Flow 4-1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-1
Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-1
Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-2
Step 1: Grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-3
Step 2: Units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-5
Step 3: Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-6
ii
CONTENTS
Step 4: Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-8

Step 7: Solution: Steady Flow . . . . . . . . . . . . . . . . . . . . . . . . 4-12
Step 8: Enable Time Dependence and Set Unsteady Conditions . . . . . 4-24
Step 9: Solution: Unsteady Flow . . . . . . . . . . . . . . . . . . . . . . 4-27
Step 10: Saving and Postprocessing Time-Dependent Data Sets . . . . . 4-30
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-43
5 Modeling Radiation and Natural Convection 5-1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-1
Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-1
Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-2
Step 1: Grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-3
Step 2: Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-6
Step 3: Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-9
Step 5: Solution for the Rosseland Model . . . . . . . . . . . . . . . . . . 5-13
Step 6: Postprocessing for the Rosseland Model . . . . . . . . . . . . . . 5-16
Step 7: P-1 Model Definition, Solution, and Postprocessing . . . . . . . . 5-25
Step 8: DTRM Definition, Solution, and Postprocessing . . . . . . . . . . 5-29
Step 9: DO Model Definition, Solution, and Postprocessing . . . . . . . . 5-33
Step 10: Comparison of y-Velocity Plots . . . . . . . . . . . . . . . . . . 5-36
Step 11: Comparison of Radiation Models for an Optically
Thick Medium . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-38
Step 12: S2S Model Definition, Solution and Postprocessing for a
Non-Participating Medium . . . . . . . . . . . . . . . . . . . . . 5-40

c Fluent Inc. January 11, 2005 iii
CONTENTS
Step 13: Comparison of Radiation Models for a Non-Participating

Medium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-45
Step 14: S2S Model Definition, Solution and Postprocessing with
Partial Enclosure . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-47
Step 15: Comparison of S2S Models with and without Partial Enclosure . 5-52
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-53
6 Using a Non-Conformal Mesh 6-1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-1
Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-1
Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-3
Step 1: Merging the Mesh Files . . . . . . . . . . . . . . . . . . . . . . . 6-4
Step 2: Grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-5
Step 3: Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-8
Step 4: Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-10
Step 7: Grid Interfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-20
Step 8: Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-21
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-31
7 Using a Single Rotating Reference Frame 7-1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-1
Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-1
Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-3
Step 1: Grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-3
iv
CONTENTS
Step 2: Units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-5

Step 3: Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-6
Step 4: Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-8
Step 6: Solution Using the Standard k- Model . . . . . . . . . . . . . . 7-12
Step 7: Postprocessing for the Standard k- Solution . . . . . . . . . . . 7-18
Step 8: Solution Using the RNG k- Model . . . . . . . . . . . . . . . . . 7-24
Step 9: Postprocessing for the RNG k- Solution . . . . . . . . . . . . . . 7-26
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-29
Further Improvements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-30
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-30
8 Using Multiple Rotating Reference Frames 8-1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-1
Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-1
Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-3
Step 1: Grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-3
Step 2: Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-6
Step 3: Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-8
Step 5: Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-15
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-22

c Fluent Inc. January 11, 2005 v
CONTENTS
9 Using the Mixing Plane Model 9-1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-1
Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-1
Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-2
Step 1: Grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-3
Step 2: Units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-5
Step 3: Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-6
Step 4: Mixing Plane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-8
Step 5: Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-10
Step 7: Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-21
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-33
10 Using Sliding Meshes 10-1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-1
Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-1
Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-3
Step 1: Merging the Mesh Files . . . . . . . . . . . . . . . . . . . . . . . 10-4
Step 2: Grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-5
Step 3: Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-8
Step 4: Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-10
vi
CONTENTS
Step 8: Solution: Steady Flow with Non-Moving Rotor . . . . . . . . . . 10-19

Step 9: Enable Time Dependence and Sliding Rotor Motion . . . . . . . 10-30
Step 10: Solution: Unsteady Flow with Moving Rotor . . . . . . . . . . . 10-33
Step 11: Postprocessing at t = 0.1 Second . . . . . . . . . . . . . . . . . 10-41
Step 12: Saving and Postprocessing Time-Dependent Data Sets . . . . . 10-45
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-49
11 Using Dynamic Meshes 11-1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-1
Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-1
Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-2
Step 1: Grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-3
Step 2: Units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-5
Step 3: Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-6
Step 4: Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-8
Step 8: Mesh Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-13
Step 9: Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-20
Step 10: Postprocessing . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-28
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-30
12 Modeling Species Transport and Gaseous Combustion 12-1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-1
Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-1
Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-3

c Fluent Inc. January 11, 2005 vii
CONTENTS

Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-3
Step 1: Grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-4
Step 2: Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-6
Step 3: Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-10
Step 5: Initial Solution Using Constant Heat Capacity . . . . . . . . . . 12-20
Step 6: Solution Using Non-Constant Heat Capacity . . . . . . . . . . . 12-25
Step 8: NOx Prediction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-38
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-47
Further Improvements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-47
13 Using the Non-Premixed Combustion Model 13-1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-1
Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-1
Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-3
Step 1: Grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-3
Step 2: Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-5
Step 3: Non Adiabatic PDF Table . . . . . . . . . . . . . . . . . . . . . . 13-8
Step 4: Models: Discrete Phase . . . . . . . . . . . . . . . . . . . . . . . 13-16
Step 5: Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-22
Step 8: Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-31
Step 10: Energy Balances and Particle Reporting . . . . . . . . . . . . . 13-41
viii
CONTENTS
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-44
Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-45
Coal Analysis for Elemental Composition . . . . . . . . . . . . . . . . . . 13-45
Discrete Phase Material Properties . . . . . . . . . . . . . . . . . . . . . 13-46
14 Modeling Surface Chemistry 14-1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-1
Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-1
Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-3
Step 1: Grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-4
Step 2: Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-7
Step 3: Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-10
Step 6: Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-23
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-32
15 Modeling Evaporating Liquid Spray 15-1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-1
Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-1
Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-2
Step 1: Grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-3
Step 2: Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-8
Step 4: Initial Solution Without Droplets . . . . . . . . . . . . . . . . . . 15-16

c Fluent Inc. January 11, 2005 ix
CONTENTS
Step 5: Create a Spray Injection . . . . . . . . . . . . . . . . . . . . . . . 15-24

Step 6: Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-30
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-37
16 Using the VOF Model 16-1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-1
Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-1
Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-2
Step 1: Grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-3
Step 2: Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-5
Step 3: Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-8
Step 4: Phases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-9
Step 7: Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-16
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-41
17 Modeling Cavitation 17-1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-1
Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-1
Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-2
Step 1: Grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-3
Step 2: Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-5
Step 3: Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-9
x
CONTENTS
Step 4: Phases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-11

Step 7: Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-17
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-24
18 Using the Mixture and Eulerian Multiphase Models 18-1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-1
Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-1
Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-2
Step 1: Grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-3
Step 2: Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-5
Step 3: Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-9
Step 4: Phases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-10
Step 6: Solution Using the Mixture Model . . . . . . . . . . . . . . . . . 18-18
Step 7: Postprocessing for the Mixture Solution . . . . . . . . . . . . . . 18-20
Step 8: Setup and Solution for the Eulerian Model . . . . . . . . . . . . . 18-24
Step 9: Postprocessing for the Eulerian Model . . . . . . . . . . . . . . . 18-28
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-30
19 Using the Eulerian Multiphase Model for Granular Flow 19-1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-1
Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-1
Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-2

c Fluent Inc. January 11, 2005 xi
CONTENTS
Step 1: Grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-3

Step 2: Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-7
Step 3: Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-10
Step 4: Phases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-12
Step 6: Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-19
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-33
20 Modeling Solidification 20-1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-1
Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-1
Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-3
Step 1: Grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-3
Step 2: Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-5
Step 3: Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-8
Step 5: Solution: Steady Conduction . . . . . . . . . . . . . . . . . . . . 20-16
Step 6: Solution: Unsteady Flow and Heat Transfer . . . . . . . . . . . . 20-23
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-31
21 Using the Eulerian Granular Multiphase Model with Heat Transfer 21-1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-1
Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-1
Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-2
Step 1: Grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-3
xii
CONTENTS
Step 2: Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-5

Step 3: Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-8
Step 4: Phases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-11
Step 6: Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-21
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-32
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-32
22 Postprocessing 22-1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-1
Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-1
Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-2
Step 1: Grid Display . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-3
Step 2: Adding Lights . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-5
Step 3: Creating Isosurfaces . . . . . . . . . . . . . . . . . . . . . . . . . 22-9
Step 4: Contours . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-10
Step 5: Velocity Vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-13
Step 6: Animation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-18
Step 7: Pathlines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-22
Step 8: Overlaying Velocity Vectors on the Pathline Display . . . . . . . 22-28
Step 9: Exploded Views . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-31
Step 10: Animating the Display of Results in Successive
Streamwise Planes . . . . . . . . . . . . . . . . . . . . . . . . . . 22-36
Step 11: XY Plots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-38
Step 12: Annotation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-42
Step 13: Saving Hardcopy Files . . . . . . . . . . . . . . . . . . . . . . . 22-44
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22-44

c Fluent Inc. January 11, 2005 xiii
CONTENTS
23 Turbo Postprocessing 23-1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-1
Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-1
Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-2
Step 1: Reading the Case and Data Files . . . . . . . . . . . . . . . . . . 23-2
Step 2: Grid Display . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-3
Step 3: Defining the Turbomachinery Topology . . . . . . . . . . . . . . 23-5
Step 4: Isosurface Creation . . . . . . . . . . . . . . . . . . . . . . . . . . 23-7
Step 5: Contours . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-9
Step 6: Reporting Turbo Quantities . . . . . . . . . . . . . . . . . . . . . 23-14
Step 7: Averaged Contours . . . . . . . . . . . . . . . . . . . . . . . . . . 23-15
Step 8: 2D Contours . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-17
Step 9: Averaged XY Plots . . . . . . . . . . . . . . . . . . . . . . . . . 23-19
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-21
24 Parallel Processing 24-1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24-1
Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24-1
Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24-3
Step 1: Starting the Parallel Version of FLUENT . . . . . . . . . . . . . . 24-3
Step 1A: Multiprocessor UNIX Machine . . . . . . . . . . . . . . . . . . 24-4
Step 1B: Multiprocessor Windows Machine . . . . . . . . . . . . . . . . . 24-6
Step 1C: Network of UNIX Workstations . . . . . . . . . . . . . . . . . . 24-7
Step 1D: Network of Windows Workstations . . . . . . . . . . . . . . . . 24-11
Step 2: Reading and Partitioning the Grid . . . . . . . . . . . . . . . . . 24-12
xiv
CONTENTS
Step 3: Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24-18

Step 4: Checking Parallel Performance . . . . . . . . . . . . . . . . . . . 24-19
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24-23

c Fluent Inc. January 11, 2005 xv
CONTENTS
xvi
Tutorial 1. Introduction to Using FLUENT: Fluid Flow and
Heat Transfer in a Mixing Elbow
Introduction
This tutorial illustrates the setup and solution of the two-dimensional turbulent fluid flow
and heat transfer in a mixing junction. The mixing elbow configuration is encountered
in piping systems in power plants and process industries. It is often important to predict
the flow field and temperature field in the neighborhood of the mixing region in order to
properly design the location of inlet pipes.
In this tutorial you will learn how to:
• Read an existing grid file into FLUENT
• Use mixed units to define the geometry and fluid properties
• Set material properties and boundary conditions for a turbulent forced convection
problem
• Initiate the calculation with residual plotting
• Calculate a solution using the segregated solver
• Examine the flow and temperature fields using graphics
• Enable the second-order discretization scheme for improved prediction of tempera-

ture
• Adapt the grid based on the temperature gradient to further improve the prediction
of temperature
Prerequisites
This tutorial assumes that you have little experience with FLUENT, but that you are
generally familiar with the interface.

c Fluent Inc. January 11, 2005 1-1
Introduction to Using FLUENT: Fluid Flow and Heat Transfer in a Mixing Elbow
Problem Description
The problem to be considered is shown schematically in Figure 1.1. A cold fluid at 26◦ C
enters through the large pipe and mixes with a warmer fluid at 40◦ C in the elbow. The
pipe dimensions are in inches, and the fluid properties and boundary conditions are given
in SI units. The Reynolds number at the main inlet is 2.03 × 105 , so that a turbulent
model will be necessary.
Figure 1.1: Problem Specification
1-2
Setup and Solution

Preparation
1. Download introduction.zip from the Fluent Inc. User Services Center
(www.fluentusers.com) to your working directory. This file can be found from
the Documentation link on the FLUENT product page.
OR,
Copy introduction.zip from the FLUENT documentation CD to your working
directory.
For UNIX systems, you can find the file by inserting the CD into your CD-ROM
drive and going to the following directory:
/cdrom/fluent6.2/help/tutfiles/
where cdrom must be replaced by the name of your CD-ROM drive.

For Windows systems, you can find the file by inserting the CD into your CD-ROM
drive and going to the following directory:
cdrom:\fluent6.2\help\tutfiles\
where cdrom must be replaced by the name of your CD-ROM drive (e.g., E).
2. Unzip introduction.zip.
elbow.msh can be found in the /introduction folder created after unzipping the
file.
3. Start the 2D version of FLUENT.

Step 1: Grid
1. Read the grid file elbow.msh.
File −→ Read −→Case...
(a) Select the file elbow.msh by clicking on it under Files and then clicking on OK.
Note: As this grid is read by FLUENT, messages will appear in the console window
reporting the progress of the conversion. After reading the grid file, FLUENT
will report that 918 triangular fluid cells have been read, along with a number
of boundary faces with different zone identifiers.
2. Check the grid.

Grid −→Check
1-4
Grid Check
Domain Extents:
x-coordinate: min (m) = 0.000000e+00, max (m) = 6.400001e+01
y-coordinate: min (m) = -4.538534e+00, max (m) = 6.400000e+01
Volume statistics:
minimum volume (m3): 2.782193e-01
maximum volume (m3): 3.926232e+00
total volume (m3): 1.682930e+03
Face area statistics:
minimum face area (m2): 8.015718e-01
maximum face area (m2): 4.118252e+00
Checking number of nodes per cell.
Checking number of faces per cell.
Checking thread pointers.
Checking number of cells per face.
Checking face cells.
Checking bridge faces.
Checking right-handed cells.
Checking face handedness.
Checking element type consistency.
Checking boundary types:
Checking face pairs.
Checking periodic boundaries.
Checking node count.
Checking nosolve cell count.
Checking nosolve face count.
Checking face children.
Checking cell children.
Checking storage.
Done.
Note: The minimum and maximum values may vary slightly when running on
different platforms. The grid check lists the minimum and maximum x and
y values from the grid, in the default SI units of meters, and reports on a
number of other grid features that are checked. Any errors in the grid would
be reported at this time. In particular, you should always make sure that the
minimum volume is not negative, since FLUENT cannot begin a calculation if
this is the case. To scale the grid to the correct units of inches, the Scale Grid
panel will be used.

3. Smooth (and swap) the grid.

Grid −→ Smooth/Swap...
To ensure the best possible grid quality for the calculation, it is good practice to
smooth a triangular or tetrahedral grid after you read it into FLUENT.
(a) Click the Smooth button and then click Swap repeatedly until FLUENT reports
that zero faces were swapped.
If FLUENT cannot improve the grid by swapping, no faces will be swapped.
(b) Close the panel.
4. Scale the grid.

Grid −→Scale...
(a) Under Units Conversion, select in from the drop-down list to complete the
phrase Grid Was Created In in (inches).
(b) Click Scale to scale the grid.
The reported values of the Domain Extents will be reported in the default SI
units of meters.
(c) Click Change Length Units to set inches as the working units for length.
Confirm that the maximum x and y values are 64 inches (see Figure 1.1).
1-6
(d) The grid is now sized correctly, and the working units for length have been
set to inches. Close the panel.
Note: Because the default SI units will be used for everything but the length, there
will be no need to change any other units in this problem. The choice of inches
for the unit of length has been made by the actions you have just taken. If you
want to change the working units for length to something other than inches,
say, mm, you would have to visit the Set Units panel in the Define pull-down
menu.
5. Display the grid (Figure 1.2).

Display −→Grid...
(a) Make sure that all of the surfaces are selected and click Display.

Grid
FLUENT 6.2 (2d, segregated, lam)
Figure 1.2: The Triangular Grid for the Mixing Elbow
Extra: You can use the right mouse button to check which zone number corresponds to
each boundary. If you click the right mouse button on one of the boundaries in the
graphics window, its zone number, name, and type will be printed in the FLUENT
console window. This feature is especially useful when you have several zones of
the same type and you want to distinguish between them quickly.
1-8
Step 2: Models
1. Keep the default solver settings.
Define −→ Models −→Solver...
2. Turn on the standard k- turbulence model.

Define −→ Models −→Viscous...
(a) Select k-epsilon in the Model list.
The original Viscous Model panel will expand when you do so.
(b) Accept the default Standard model by clicking OK.

3. Enable heat transfer by activating the energy equation.

Define −→ Models −→Energy...
1-10
Step 3: Materials
1. Create a new material called water.
Define −→Materials...
(a) Type the name water in the Name text-entry box.
(b) Enter the values shown in the table below under Properties:
Property Value
density 1000 kg/m3
Cp 4216 J/kg-K
thermal conductivity 0.677 W/m-K
viscosity 8 ×10−4 kg/m-s
(c) Click Change/Create.

(d) Click No when FLUENT asks if you want to overwrite air.

The material water will be added to the list of materials which originally con-
tained only air. You can confirm that there are now two materials defined by
examining the drop-down list under Fluid Materials.
Extra: You could have copied the material water from the materials database
(accessed by clicking on the Fluent Database... button). If the properties
in the database are different from those you wish to use, you can still edit
the values under Properties and click the Change/Create button to update
your local copy. (The database will not be affected.)
(e) Close the Materials panel.
1-12
Step 4: Boundary Conditions

1. Set the conditions for the fluid.

(a) Select fluid-9 under Zone.
The Type will be reported as fluid.
(b) Click Set... to open the Fluid panel.
(c) Specify water as the fluid material by selecting water in the Material Name
drop-down list. Click on OK.

2. Set the boundary conditions at the main inlet.

(a) Select velocity-inlet-5 under Zone and click Set....
Hint: If you are unsure of which inlet zone corresponds to the main inlet, you
can probe the grid display with the right mouse button and the zone ID will
be displayed in the FLUENT console window. In the Boundary Conditions
panel, the zone that you probed will automatically be selected in the Zone
list. In 2D simulations, it may be helpful to return to the Grid Display
panel and deselect the display of the fluid and interior zones (in this case,
internal-3) before probing with the mouse button for zone names.
1-14
(b) Choose Components as the Velocity Specification Method.

(c) Set an X-Velocity of 0.2 m/s.
(d) Retain Y-Velocity at 0 m/s.
(e) Set a Temperature of 293 K.
(f) Select Intensity and Hydraulic Diameter as the Turbulence Specification Method.
(g) Enter a Turbulence Intensity of 5%, and a Hydraulic Diameter of 32 in.
The hydraulic diameter Dh is defined as:
4A
Dh = ,
Pw
where A is the cross-sectional area and Pw is the wetted perimeter. In this
2D case, the wetted perimeter for a unit depth slice is equal to 2, since we are
modelling a unit depth slice of a 3D duct that is far removed from the walls
on either side.
3. Repeat this operation for velocity-inlet-6, using the values in the following table:
velocity specification method components
y velocity 1.0 m/s
x velocity 0 m/s
temperature 313 K
turbulence specification method intensity & hydraulic diameter
turbulence intensity 5%
hydraulic diameter 8 in

4. Set the boundary conditions for pressure-outlet-7, as shown in the panel below.
These values will be used in the event that flow enters the domain through this
boundary.
5. For wall-4, keep the default settings for a Heat Flux of 0.
1-16
6. For wall-8, you will also keep the default settings.

Note: If you probe your display of the grid (without the interior cells) you will see
that wall-8 is the wall on the outside of the bend just after the junction. This
separate wall zone has been created for the purpose of doing certain postpro-
cessing tasks, to be discussed later in this tutorial.

Step 5: Solution
1. Initialize the flow field using the boundary conditions set at velocity-inlet-5.
Solve −→ Initialize −→Initialize...
(a) Choose velocity-inlet-5 from the Compute From list.
(b) Add a Y Velocity value of 0.2 m/sec throughout the domain.
Note: While an initial X Velocity is an appropriate guess for the horizontal
section, the addition of a Y Velocity will give rise to a better initial guess
throughout the entire elbow.
(c) Click Init and Close the panel.
1-18
2. Enable the plotting of residuals during the calculation.

Solve −→ Monitors −→Residual...
(a) Select Plot under Options, and click OK.

Note: By default, all variables will be monitored and checked for determining the
convergence of the solution. Although residuals are used for checking conver-
gence, a more reliable method is to define a surface monitor.

3. Define a surface monitor.

Solve −→ Monitors −→Surface...
(a) Increase the number of Surface Monitors to 1.

(b) Enable Plot and Write.
(c) Click Define... to open the Define Surface Monitor panel.
i. Under Report of, select Temperature... and Static Temperature.

ii. Under Report Type, select Mass-Weighted Average.
iii. Under Surfaces, select pressure-outlet-7.
iv. Click OK.
1-20
4. Save the case file (elbow1.cas).

File −→ Write −→Case...
Keep the Write Binary Files (default) option on so that a binary file will be written.
5. Start the calculation by requesting 100 iterations.

Solve −→Iterate...
(a) Input 100 for the Number of Iterations and click Iterate.
The solution reaches convergence after approximately 55 iterations. The resid-

ual plot and the convergence history of mass-weighted average temperature are
shown in Figures 1.3 and 1.4, respectively.

Note: The number of iterations required for convergence varies according to

the platform used. Also, since the residual values are different for different
computers, the plot that appears on your screen may not be exactly the
same as the one shown here.
Residuals
continuity
x-velocity
1e+03
y-velocity
energy
k 1e+02
epsilon
1e+01
1e+00
1e-01
1e-02
1e-03
1e-04
1e-05
1e-06
1e-07
0 10 20 30 40 50 60
Iterations
Scaled Residuals
FLUENT 6.2 (2d, segregated, ske)
Figure 1.3: Residuals for the First 60 Iterations
6. Check for convergence.

There are no universal metrics for judging convergence. Residual definitions that
are useful for one class of problem are sometimes misleading for other classes of
problems. Therefore it is a good idea to judge convergence not only by examining
residual levels, but also by monitoring relevant integrated quantities and checking
for mass and energy balances.
The three methods to check for convergence are:
• Monitoring the residuals.
Convergence will occur when the Convergence Criterion for each variable has
been reached. The default criterion is that each residual will be reduced to
a value of less than 10−3 , except the energy residual, for which the default
criterion is 10−6 .
• Solution no longer changes with more iterations.
Sometimes the residuals may not fall below the convergence criterion set in
the case setup. However, monitoring the representative flow variables through
iterations may show that the residuals have stagnated and do not change with
further iterations. This could also be considered as convergence.
1-22
Monitors
monitor-1
310.0000
308.0000
306.0000
304.0000
302.0000
Mass 300.0000
Weighted
298.0000
Average
(k) 296.0000
294.0000
292.0000
290.0000
0 10 20 30 40 50 60
Iteration
Convergence history of Static Temperature on pressure-outlet-7

Figure 1.4: Convergence History of Mass-Weighted Average Temperature
• Overall mass, momentum, energy and scalar balances are obtained.

Check the overall mass, momentum, energy and scalar balances in the Flux
Reports panel. The net imbalance should be less than 0.2% of the net flux
through the domain.
Report −→Fluxes

(a) In the Boundaries list, select pressure-outlet-7, velocity-inlet-5, and velocity-

inlet-6.
(b) Click Compute.
7. Save the data file (elbow1.dat).

Use the same prefix (elbow1) that you used when you saved the case file earlier.
Note that additional case and data files will be written later in this session.
File −→ Write −→Data...
1-24
Step 6: Displaying the Preliminary Solution

1. Display filled contours of velocity magnitude (Figure 1.5).
Display −→ Contours...
(a) Select Velocity... and then Velocity Magnitude from the drop-down lists under
Contours of.
(b) Select Filled under Options.
(c) Click Display.
Note: Right-clicking on a point in the domain will cause the value of the corre-
sponding contour to be displayed in the console window.

1.24e+00
1.18e+00
1.12e+00
1.05e+00
9.93e-01
9.31e-01
8.69e-01
8.07e-01
7.45e-01
6.82e-01
6.20e-01
5.58e-01
4.96e-01
4.34e-01
3.72e-01
3.10e-01
2.48e-01
1.86e-01
1.24e-01
6.20e-02
0.00e+00
Contours of Velocity Magnitude (m/s)

Figure 1.5: Predicted Velocity Distribution After the Initial Calculation
1-26
2. Display filled contours of temperature (Figure 1.6).
(a) Select Temperature... and Static Temperature in the drop-down lists under
Contours of.
(b) Click Display.

3.13e+02
3.12e+02
3.11e+02
3.10e+02
3.09e+02
3.08e+02
3.07e+02
3.06e+02
3.05e+02
3.04e+02
3.03e+02
3.02e+02
3.01e+02
3.00e+02
2.99e+02
2.98e+02
2.97e+02
2.96e+02
2.95e+02
2.94e+02
2.93e+02
Contours of Static Temperature (k)

Figure 1.6: Predicted Temperature Distribution After the Initial Calculation
1-28
3. Display velocity vectors (Figure 1.7).

Display −→ Vectors...
(a) Click Display to plot the velocity vectors.
Note: The Auto Scale button is on by default under Options. This scaling
sometimes creates vectors that are too small or too large in the majority
of the domain.
(b) Resize the vectors by increasing the Scale factor to 3.
(c) Display the vectors once again.

(d) Use the middle mouse button to zoom the view. To do this, hold down the
button and drag your mouse to the right and either up or down to construct
a rectangle on the screen. The rectangle should be a frame around the region
that you wish to enlarge. Let go of the mouse button and the image will be
redisplayed (Figure 1.8).
(e) Un-zoom the view by holding down the middle mouse button and dragging it
to the left to create a rectangle. When you let go, the image will be redrawn.
If the resulting image is not centered, you can use the left mouse button to
translate it on your screen.

1.40e+00
1.33e+00
1.27e+00
1.20e+00
1.13e+00
1.06e+00
9.96e-01
9.28e-01
8.61e-01
7.94e-01
7.26e-01
6.59e-01
5.91e-01
5.24e-01
4.56e-01
3.89e-01
3.22e-01
2.54e-01
1.87e-01
1.19e-01
5.19e-02
Velocity Vectors Colored By Velocity Magnitude (m/s)

Figure 1.7: Resized Velocity Vectors
1.40e+00
1.33e+00
1.27e+00
1.20e+00
1.13e+00
1.06e+00
9.96e-01
9.28e-01
8.61e-01
7.94e-01
7.26e-01
6.59e-01
5.91e-01
5.24e-01
4.56e-01
3.89e-01
3.22e-01
2.54e-01
1.87e-01
1.19e-01
5.19e-02

Figure 1.8: Magnified View of Velocity Vectors
1-30
4. Create an XY plot of temperature across the exit (Figure 1.9).

Plot −→ XY Plot...
(a) Select Temperature... and Static Temperature in the drop-down lists under the
Y Axis Function.
(b) Select pressure-outlet-7 from the Surfaces list.
(c) Click Plot.

pressure-outlet-7
3.10e+02
3.08e+02
3.06e+02
3.04e+02
Static
Temperature 3.02e+02
(k)
3.00e+02
2.98e+02
2.96e+02
48 50 52 54 56 58 60 62 64
Position (in)
Static Temperature
Figure 1.9: Temperature Distribution at the Outlet
1-32
5. Make an XY plot of the static pressure on the outer wall of the large pipe, wall-8
(Figure 1.10).
(a) Choose Pressure... and Static Pressure from the Y Axis Function drop-down
lists.
(b) Deselect pressure-outlet-7 and select wall-8 from the Surfaces list.
(c) Change the Plot Direction for X to 0, and the Plot Direction for Y to 1.
With a Plot Direction vector of (0,1), FLUENT will plot static pressure at the
cells of wall-8 as a function of y.
(d) Click Plot.

wall-8
1.00e+02
0.00e+00
-1.00e+02
-2.00e+02
Static
Pressure -3.00e+02
(pascal)
-4.00e+02
-5.00e+02
-6.00e+02
10 20 30 40 50 60 70
Position (in)
Static Pressure
Figure 1.10: Pressure Distribution along the Outside Wall of the Bend
1-34
6. Define a custom field function for the dynamic head formula (ρ|V |2 /2).
Define −→ Custom Field Functions...
(a) In the Field Functions drop-down list, select Density and click the Select button.
(b) Click the multiplication button, X.
(c) In the Field Functions drop-down list, select Velocity and Velocity Magnitude
and click Select.
(d) Click y^x to raise the last entry to a power, and click 2 for the power.
(e) Click the divide button, /, and then click 2.
(f) Enter the name dynam-head in the New Function Name text entry box.
(g) Click Define, and then Close the panel.

7. Display filled contours of the custom field function (Figure 1.11).

(a) Select Custom Field Functions... in the drop-down list under Contours of.
The function you created, dynam-head, will be shown in the lower drop-down
list.
(b) Click Display, and then Close the panel.
Note: You may need to un-zoom your view after the last vector display, if you have
not already done so.
8. Write the case and data files to save the settings for the custom field function.
File −→ Write −→Case & Data...
1-36
7.70e+02
7.31e+02
6.93e+02
6.54e+02
6.16e+02
5.77e+02
5.39e+02
5.00e+02
4.62e+02
4.23e+02
3.85e+02
3.46e+02
3.08e+02
2.69e+02
2.31e+02
1.92e+02
1.54e+02
1.15e+02
7.70e+01
3.85e+01
0.00e+00
Contours of dynam-head
Figure 1.11: Contours of the Custom Field Function, Dynamic Head

Step 7: Enabling Second-Order Discretization

The elbow solution computed in the first part of this tutorial uses first-order discretiza-
tion. The resulting solution is very diffusive; mixing is overpredicted, as can be seen
in the contour plots of temperature and velocity distribution. You will now change to
second-order discretization for all listed equations, in order to improve the accuracy of
the solution. With the second-order discretization, you will change the gradient option in
the solver from cell-based to node-based in order to optimize energy conservation.
1. Change the Gradient Option in the Solver panel.

Define −→ Models −→ Solver...
(a) Under Gradient Option, select Node-Based.
This option is more suitable than the cell-based gradient option for meshes
with tri-elements (Figure 1.2), as it will ensure better energy conservation.
2. Enable the second-order scheme for the calculation of all the listed equations.
Solve −→ Controls −→Solution...
(a) Under Discretization, select Second Order for Pressure, Second Order Upwind for
Momentum, Turbulence Kinetic Energy, Turbulence Dissipation Rate, and Energy.
(b) Keep the default Under-Relaxation Factors settings.
Note: You will have to scroll down both the Discretization and Under-Relaxation
Factors lists to see the Energy options.
1-38
(c) Click OK.
3. Continue the calculation by requesting 100 more iterations.

Solve −→ Iterate...
To save the convergence history for this set of iterations as a separate output file,
you can change the File Name in the Define Surface Monitor to monitor-2.out.
The solution converges in approximately 50 additional iterations (Figure 1.12). The
convergence history is shown in Figure 1.13.
Residuals
continuity
x-velocity
1e+03
y-velocity
energy
k 1e+02
epsilon
1e+01
1e+00
1e-01
1e-02
1e-03
1e-04
1e-05
1e-06
1e-07
0 20 40 60 80 100 120
Iterations
Scaled Residuals
Figure 1.12: Residuals for the Second-Order Energy Calculation
Note: Whenever you change the solution control parameters, it is natural to see
the residuals jump.

Monitors
monitor-1
304.4000
304.3000
304.2000
304.1000
Mass 304.0000
Weighted
Average 303.9000
(k)
303.8000
303.7000
303.6000
50 60 70 80 90 100 110
Iteration

1-40
4. Write the case and data files for the second-order solution (elbow2.cas and elbow2.dat).
File −→ Write −→ Case & Data...
(a) Enter the name elbow2 in the Case/Data File box.
(b) Click OK.
The files elbow2.cas and elbow2.dat will be created in your directory.
5. Examine the revised temperature distribution (Figure 1.14).

The thermal spreading after the elbow has been reduced from the earlier prediction
(Figure 1.6).

3.13e+02
3.12e+02
3.11e+02
3.10e+02
3.09e+02
3.08e+02
3.07e+02
3.06e+02
3.05e+02
3.04e+02
3.03e+02
3.02e+02
3.01e+02
3.00e+02
2.99e+02
2.98e+02
2.97e+02
2.96e+02
2.95e+02
2.94e+02
2.93e+02

Figure 1.14: Temperature Contours for the Second-Order Solution
1-42
Step 8: Adapting the Grid

The elbow solution can be improved further by refining the grid to better resolve the flow
details. In this step, you will adapt the grid based on the temperature gradients in the
current solution. Before adapting the grid, you will first determine an acceptable range
of temperature gradients over which to adapt. Once the grid has been refined, you will
continue the calculation.
1. Plot filled contours of temperature on a cell-by-cell basis (Figure 1.15).

(a) Select Temperature... and Static Temperature in the Contours of drop-down

lists.
(b) Deselect Node Values under Options and click Display.
Note: When the contours are displayed you will see the cell values of temper-
ature instead of the smooth-looking node values. Node values are obtained
by averaging the values at all of the cells that share the node. Cell val-
ues are the values that are stored at each cell center and are displayed
throughout the cell. Examining the cell-by-cell values is helpful when you
are preparing to do an adaption of the grid because it indicates the re-
gion(s) where the adaption will take place.

2. Plot the temperature gradients that will be used for adaption (Figure 1.16).
(a) Select Adaption... and Adaption Function in the Contours of drop-down lists.
(b) Click Display to see the gradients of temperature, displayed on a cell-by-cell
basis.
3.13e+02
3.12e+02
3.11e+02
3.10e+02
3.09e+02
3.08e+02
3.07e+02
3.06e+02
3.05e+02
3.04e+02
3.03e+02
3.02e+02
3.01e+02
3.00e+02
2.99e+02
2.98e+02
2.97e+02
2.96e+02
2.95e+02
2.94e+02
2.93e+02

Figure 1.15: Temperature Contours for the Second-Order Solution: Cell Values
1-44
1.25e-01
1.19e-01
1.13e-01
1.06e-01
1.00e-01
9.39e-02
8.76e-02
8.14e-02
7.51e-02
6.88e-02
6.26e-02
5.63e-02
5.01e-02
4.38e-02
3.75e-02
3.13e-02
2.50e-02
1.88e-02
1.25e-02
6.26e-03
1.42e-14
Contours of Adaption Function

Figure 1.16: Contours of Adaption Function: Temperature Gradient
Note: The quantity Adaption Function defaults to the gradient of the variable
whose Max and Min were most recently computed in the Contours panel.
In this example, the static temperature is the most recent variable to have
its Max and Min computed, since this occurs when the Display button is
pushed. Note that for other applications, gradients of another variable
might be more appropriate for performing the adaption.
3. Plot temperature gradients over a limited range in order to mark cells for adaption
(Figure 1.17).
(a) Under Options, deselect Auto Range so that you can change the minimum
temperature gradient value to be plotted.
The Min temperature gradient is 0 K/m, as shown in the Contours panel.
(b) Enter a new Min value of 0.02.
(c) Click Display.
The colored cells in the figure are in the “high gradient” range, so they will be
the ones targeted for adaption.

1.25e-01
1.20e-01
1.15e-01
1.09e-01
1.04e-01
9.89e-02
9.36e-02
8.84e-02
8.31e-02
7.78e-02
7.26e-02
6.73e-02
6.21e-02
5.68e-02
5.15e-02
4.63e-02
4.10e-02
3.58e-02
3.05e-02
2.53e-02
2.00e-02
Contours of Adaption Function

Figure 1.17: Contours of Temperature Gradient Over a Limited Range
1-46
4. Adapt the grid in the regions of high temperature gradient.

Adapt −→ Gradient...
(a) Select Temperature... and Static Temperature in the Gradients of drop-down
lists.
(b) Deselect Coarsen under Options, so that only a refinement of the grid will be
performed.
(c) Click Compute.
FLUENT will update the Min and Max values.
(d) Enter the value of 0.02 for the Refine Threshold.

(e) Click Mark.

FLUENT will report the number of cells marked for adaption in the console
window.
(f) Click Manage... to display the marked cells.
This will open the Manage Adaption Registers panel.
(g) Click Display.

FLUENT will display the cells marked for adaption (Figure 1.18).
(h) Click Adapt. Click Yes when you are asked for confirmation.
Note: There are two different ways to adapt. You can click on Adapt in the
Manage Adaption Registers panel as was just done, or Close this panel and
do the adaption in the Gradient Adaption panel. If you use the Adapt
button in the Gradient Adaption panel, FLUENT will recreate an adaption
register. Therefore, once you have the Manage Adaption Registers panel
open, it saves time to use the Adapt button there.
(i) Close the Manage Adaption Registers and Gradient Adaption panels.
1-48
Adaption Markings (gradient-r0)

Figure 1.18: Cells Marked for Adaption

5. Display the adapted grid (Figure 1.19).

Display −→ Grid...
Grid
Figure 1.19: The Adapted Grid
6. Request an additional 100 iterations.

Solve −→ Iterate...
The solution converges after approximately 50 additional iterations.
7. Write the final case and data files (elbow3.cas and elbow3.dat) using the prefix
elbow3.
File −→ Write −→ Case & Data...
1-50
Residuals
continuity
x-velocity
1e+03
y-velocity
energy
k 1e+02
epsilon
1e+01
1e+00
1e-01
1e-02
1e-03
1e-04
1e-05
1e-06
1e-07
0 20 40 60 80 100 120 140 160
Iterations
Scaled Residuals
Figure 1.20: The Complete Residual History
Monitors
monitor-1
304.2000
304.1500
304.1000
304.0500
Mass 304.0000
Weighted
Average 303.9500
(k)
303.9000
303.8500
303.8000
100 105 110 115 120 125 130 135 140 145 150
Iteration


8. Examine the filled temperature distribution (using node values) on the revised grid
(Figure 1.22).
3.13e+02
3.12e+02
3.11e+02
3.10e+02
3.09e+02
3.08e+02
3.07e+02
3.06e+02
3.05e+02
3.04e+02
3.03e+02
3.02e+02
3.01e+02
3.00e+02
2.99e+02
2.98e+02
2.97e+02
2.96e+02
2.95e+02
2.94e+02
2.93e+02

Figure 1.22: Filled Contours of Temperature Using the Adapted Grid
Summary
Comparison of the filled temperature contours for the first solution (using the original
grid and first-order discretization) and the last solution (using an adapted grid and
second-order discretization) clearly indicate that the latter is much less diffusive. While
first-order discretization is the default scheme in FLUENT, it is good practice to use
your first-order solution as a starting guess for a calculation that uses a higher-order
discretization scheme and, optionally, an adapted grid.
Note that in this problem, the flow field is decoupled from temperature since all properties
are constant. For such cases, it is more efficient to compute the flow-field solution first
(i.e., without solving the energy equation) and then solve for energy (i.e., without solving
the flow equations). You will use the Solution Controls panel to turn solution of the
equations on and off during this procedure.
1-52
Tutorial 2. Modeling Periodic Flow and Heat Transfer
Introduction
Many industrial applications, such as steam generation in a boiler or air cooling in the
coil of an air conditioner, can be modeled as two-dimensional periodic heat flow. This
tutorial illustrates how to set up and solve a periodic heat transfer problem, given a
pregenerated mesh.
The system that is modeled is a bank of tubes containing a flowing fluid at one temper-
ature that is immersed in a second fluid in cross-flow at a different temperature. Both
fluids are water, and the flow is classified as laminar and steady, with a Reynolds number
of approximately 100. The mass flow rate of the cross-flow is known, and the model is
used to predict the flow and temperature fields that result from convective heat transfer.
Due to symmetry of the tube bank and the periodicity of the flow inherent in the tube
bank geometry, only a portion of the geometry will be modeled in FLUENT, with sym-
metry applied to the outer boundaries. The resulting mesh consists of a periodic module
with symmetry. In the tutorial, the inflow boundary will be redefined as a periodic zone,
and the outflow boundary defined as its shadow.
• Create periodic zones.
• Define a specified periodic mass flow rate.
• Model periodic heat transfer with specified temperature boundary conditions.
• Calculate a solution using the segregated solver.
• Plot temperature profiles on specified isosurfaces.
Prerequisites
This tutorial assumes that you are familiar with the menu structure in FLUENT and that
you have completed Tutorial 1 . Some steps in the setup and solution procedure will not
be shown explicitly.

Modeling Periodic Flow and Heat Transfer
Problem Description
This problem considers a 2D section of a tube bank. A schematic of the problem is
shown in Figure 2.1. The bank consists of uniformly spaced tubes with a diameter of 1
cm, that are staggered in the direction of cross-fluid flow. Their centers are separated by
a distance of 2 cm in the x direction, and 1 cm in the y direction. The bank has a depth
of 1 m.
4 cm
Τ
∞ = 300 K
⋅
m = 0.05 kg/s
{ Τ wall = 400 K
0.5 cm
1 cm
3
ρ = 998.2 kg/m
µ = 0.001003 kg/m-s
c p = 4182 J/kg-K
k = 0.6 W/m-K
Figure 2.1: Schematic of the Problem
Because of the symmetry of the tube bank geometry, only a portion of the domain needs
to be modeled. The computational domain is shown in outline in Figure 2.1. A mass
flow rate of 0.05 kg/s is applied to the inflow boundary of the periodic module. The
temperature of the tube wall (Twall ) is 400 K and the bulk temperature of the cross-flow
water (T∞ ) is 300 K. The properties of water that are used in the model are shown in
Figure 2.1.
2-2
Setup and Solution

Preparation
1. Download periodic_flow_heat.zip from the Fluent Inc. User Services Center
or copy it from the FLUENT documentation CD to your working directory (as
described in Tutorial 1).
2. Unzip periodic_flow_heat.zip.
tubebank.msh can be found in the /periodic flow heat folder created after un-
zipping the file.
Step 1: Grid
1. Read the mesh file, tubebank.msh.
2. Check the grid.
Grid −→Check
FLUENT will perform various checks on the mesh and report the progress in the
console window. Make sure that the minimum volume reported is a positive num-
ber.
3. Scale the grid.
Grid −→Scale...
(a) Under Units Conversion, select cm (centimeters) from the drop-down list to
complete the phrase Grid Was Created In cm.
(b) Click Scale to scale the grid and close the panel.

4. Display the mesh (Figure 2.2).

Grid
Figure 2.2: Mesh for the Periodic Tube Bank
2-4
Quadrilateral cells are used in the regions surrounding the tube walls, and triangular
cells are used for the rest of the domain, resulting in a hybrid mesh (See Figure 2.2).
The quadrilateral cells provide better resolution of the viscous gradients near the
tube walls. The remainder of the computational domain is conveniently filled with
triangular cells.
Extra: Right-click on one of the boundaries in the graphics window to check which
zone number corresponds to the boundary. The zone number, name, and type
will be printed in the FLUENT console window. This feature is especially useful
when there are several zones of the same type and you want to distinguish
between them.
5. Create the periodic zone.

The inflow (wall-9) and outflow (wall-12) boundaries currently defined as wall zones
need to be redefined as periodic. Redefine the wall-9 boundary as translationally
periodic zone, and wall-12 as periodic shadow of wall-9.
(a) In the console window, enter the inputs shown in boxes in the following dialog.
Hint: Press <Enter> to get the command prompt (>).
grid/modify-zones/make-periodic
Periodic zone [()] 9
Shadow zone [()] 12
Rotational periodic? (if no, translational) [yes] no
Create periodic zones? [yes] yes
Auto detect translation vector? [yes] yes
computed translation deltas: 0.040000 0.000000

all 26 faces matched for zones 9 and 12.
zone 12 deleted
created periodic zones.

Step 2: Models

2-6
3. Set the periodic flow conditions.

Define −→Periodic Conditions...
(a) Under Type, select Specify Mass Flow.

This will allow you to specify the Mass Flow Rate.
(b) For Mass Flow Rate, enter a value of 0.05 and click OK.

Step 3: Materials
Add liquid water to the list of fluid materials by copying it from the materials database.
1. Copy the properties of liquid water from the database.

(a) Click on the Fluent Database... button.
This will open the Fluent Database Materials panel.
(b) Select water-liquid (h2o<l>) from the Fluent Fluid Materials list.
This will display the default settings for water-liquid. You will have to scroll
down the Fluent Fluid Materials list to see the entries.
(c) Click Copy and Close the panel.
The Materials panel will now display the copied information of water.
2-8


1. Set the conditions for fluid-16.

(a) In the Material Name drop-down list, select water-liquid
2-10
2. Set the boundary conditions for wall-21.

This is the bottom wall of the left tube in the periodic module shown in Figure 2.1.
(a) Change Zone Name to wall-bottom.

(b) Under Thermal Conditions, select Temperature.
(c) Set the Temperature to 400 K.


This is the top wall of the right tube in the periodic module shown in Figure 2.1.
(a) Change the Zone Name to wall-top.

(b) Under Thermal Conditions, select Temperature.
(d) Click OK to close the panel.
2-12
Step 5: Solution
1. Set the solution parameters.
(a) Set the Under-Relaxation Factor for Energy to 0.9.

Hint: Scroll down the Under-Relaxation Factors list to see Energy.
(b) Under Discretization, select Second Order Upwind for Momentum and Energy.

2. Enable the plotting of residuals.

(a) Under Options, enable Plot and click OK.

3. Initialize the solution.
2-14
(a) Under Initial Values, check that the value for Temperature is set to 300 K.
(b) Click Init and Close the panel.
4. Save the case file, tubebank.cas.


(a) Set the Number of Iterations to 350.

(b) Click Iterate.
The energy residual curve begins to flatten out after about 350 iterations. For the
solution to converge, you need to reduce the under-relaxation factor for energy.
6. Change the Under-Relaxation Factor for Energy to 0.6.

7. Continue the calculation by requesting another 300 iterations.

After restarting the calculation, you will see an initial dip in the plot of the energy
residual as a result of reduction in the under-relaxation factor. The solution will
converge in a total of approximately 580 iterations.
8. Save the case and data files, tubebank.cas and tubebank.dat.


Step 6: Postprocessing
1. Display filled contours of static pressure (Figure 2.3).
Display −→Contours...
(a) Under Options, select Filled.

(b) In the Contours of drop-down list, select Pressure... and Static Pressure.
(c) Click Display (Figure 2.3).
2. Change the view to mirror the display across the symmetry planes (Figure 2.4).
Display −→Views...
2-16
8.18e-02
7.55e-02
6.92e-02
6.28e-02
5.65e-02
5.02e-02
4.38e-02
3.75e-02
3.12e-02
2.48e-02
1.85e-02
1.22e-02
5.82e-03
-5.13e-04
-6.85e-03
-1.32e-02
-1.95e-02
-2.59e-02
-3.22e-02
-3.85e-02
-4.49e-02
Contours of Static Pressure (pascal)

Figure 2.3: Contours of Static Pressure
(a) Under Mirror Planes, select all of the symmetry zones by clicking the shaded
icon at the right side.
Note: There are four symmetry zones in the Mirror Planes list because the
top and bottom symmetry planes in the domain are each comprised of
two symmetry zones, one on each side of the tube. It is also possible to
generate the same display shown in Figure 2.4 by selecting just one of the
symmetry zones on the top symmetry plane, and one on the bottom.
(b) Click Apply and Close the panel.

Use the left button of your mouse to translate the view so that it is centered
in the window.
8.18e-02
7.55e-02
6.92e-02
6.28e-02
5.65e-02
5.02e-02
4.38e-02
3.75e-02
3.12e-02
2.48e-02
1.85e-02
1.22e-02
5.82e-03
-5.13e-04
-6.85e-03
-1.32e-02
-1.95e-02
-2.59e-02
-3.22e-02
-3.85e-02
-4.49e-02

Figure 2.4: Contours of Static Pressure with Symmetry
Note: The pressure contours displayed in Figure 2.4 do not include the linear
pressure gradient computed by the solver. Thus the contours are periodic at
the inflow and outflow boundaries.
2-18
3. Display filled contours of static temperature (Figure 2.5).

(a) In the Contours of drop-down list, select Temperature... and Static Temperature
and click Display (Figure 2.5).
The contours reveal the temperature increase in the fluid due to heat transfer from
the tubes. The hotter fluid is confined to the near-wall and wake regions, while a
narrow stream of cooler fluid is convected through the tube bank.

4.00e+02
3.94e+02
3.88e+02
3.82e+02
3.75e+02
3.69e+02
3.63e+02
3.57e+02
3.51e+02
3.45e+02
3.39e+02
3.32e+02
3.26e+02
3.20e+02
3.14e+02
3.08e+02
3.02e+02
2.96e+02
2.89e+02
2.83e+02
2.77e+02

Figure 2.5: Contours of Static Temperature
2-20
4. Display the velocity vectors (Figure 2.6).

Display −→Vectors...
(a) In the Color By drop-down list, select Velocity... and Velocity Magnitude.
(b) Set the Scale to 2 and click Display.
This will enlarge the displayed vectors that are displayed, making it easier to
view the flow patterns.
(c) Zoom in on the upper right portion of the left tube using your middle mouse
button, to get the display shown in Figure 2.6.
This zoomed-in view of the velocity vector plot clearly shows the recirculating
flow behind the tube and the boundary layer development along the tube surface.
5. Plot the temperature profiles at three cross sections of the tube bank.
(a) Create an isosurface on the periodic tube bank at x = 0.01 m (through the
first tube).
Create a surface of constant x coordinate for each the cross sections, x = 0.01,
0.02, and 0.03 m. These isosurfaces correspond to the vertical cross sections

1.31e-02
1.25e-02
1.18e-02
1.12e-02
1.05e-02
9.85e-03
9.19e-03
8.53e-03
7.88e-03
7.22e-03
6.56e-03
5.91e-03
5.25e-03
4.60e-03
3.94e-03
3.28e-03
2.63e-03
1.97e-03
1.31e-03
6.58e-04
1.93e-06

Figure 2.6: Velocity Vectors
through the first tube, halfway between the two tubes, and through the second
tube.
Surface −→Iso-Surface...
i. In the Surface of Constant drop-down lists, select Grid... and X-Coordinate.
ii. Enter x=0.01m under New Surface Name.
iii. Enter 0.01 for Iso-Values and click Create.
iv. Follow the same procedure to create surfaces at x = 0.02 m (halfway
between the two tubes) and x = 0.03 m (through the middle of the second
tube).
2-22
(b) Create an XY plot of static temperature on the three isosurfaces.

Plot −→XY Plot...
i. Change the Plot Direction for X to 0, and the Plot Direction for Y to 1.
With a Plot Direction vector of (0,1), FLUENT will plot the selected vari-

able as a function of y. Since you are plotting the temperature profile on

cross sections of constant x, the temperature varies with the y direction.
ii. Select Temperature... and Static Temperature in the Y-Axis Function drop-
down lists.
iii. Scroll down the Surfaces list and select x=0.01m, x=0.02m, and x=0.03m.
iv. Click Curves....
This will open the Curves - Solution XY Plot panel, used to define different
plot styles for the different plot curves.
A. Select + in the Symbol drop-down list.

B. Click Apply.
This assigns the + symbol to the x = 0.01 m curve.
C. Increase the Curve # to 1 to define the style for the x = 0.02 m curve.
D. Select x in the Symbol drop-down list.
E. Set the value of Size to 0.5.
F. Click Apply and close the panel.
Since you did not change the curve style for the x = 0.03 m curve,
the default symbol will be used.
v. In the Solution XY Plot panel, click Plot.
2-24
x=0.01m
x=0.02m
x=0.03m
4.00e+02
3.80e+02
3.60e+02
3.40e+02
Static
(k)
3.00e+02
2.80e+02
2.60e+02
0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008 0.009 0.01
Position (m)
Static Temperature
Figure 2.7: Static Temperature at x=0.01, 0.02, and 0.03 m

Summary
In this tutorial, periodic flow and heat transfer in a staggered tube bank were modeled
in FLUENT. The model was set up assuming a known mass flow through the tube bank
and constant wall temperatures. Due to the periodic nature of the flow and symmetry of
the geometry, only a small piece of the full geometry was modeled. In addition, the tube
bank configuration lent itself to the use of a hybrid mesh with quadrilateral cells around
the tubes and triangles elsewhere.
The Periodicity Conditions panel makes it easy to run this type of model over a variety
of operating conditions. For example, different flow rates (and hence different Reynolds
numbers) can be studied, or a different inlet bulk temperature can be imposed. The
resulting solution can then be examined to extract the pressure drop per tube row and
overall Nusselt number for a range of Reynolds numbers.
2-26
Tutorial 3. Modeling External Compressible Flow
Introduction
The purpose of this tutorial is to compute the turbulent flow past a transonic airfoil at
a non-zero angle of attack. You will use the Spalart-Allmaras turbulence model.
• Model compressible flow (using the ideal gas law for density)
• Set boundary conditions for external aerodynamics
• Use the Spalart-Allmaras turbulence model
• Calculate a solution using the coupled implicit solver
• Use force and surface monitors to check solution convergence
• Check the grid by plotting the distribution of y +
Prerequisites

Modeling External Compressible Flow
Problem Description
The problem considers the flow around an airfoil at an angle of attack α = 4◦ and a free
stream Mach number of 0.8 (M∞ = 0.8). This flow is transonic, and has a fairly strong
shock near the mid-chord (x/c = 0.45) on the upper (suction) side. The chord length is
1 m. The geometry of the airfoil is shown in Figure 3.1.
α = 4°
M∞= 0.8
1m
Setup and Solution

Preparation
1. Download external_compressible.zip from the Fluent Inc. User Services Center
2. Unzip external_compressible.zip.
airfoil.msh can be found in the /external compressible folder created after
unzipping the file.
3-2
Step 1: Grid
1. Read the grid file airfoil.msh.
As FLUENT reads the grid file, it will report its progress in the console window.
2. Check the grid.

Grid −→Check
FLUENT will perform various checks on the mesh and will report the progress in the
console window. Pay particular attention to the reported minimum volume. Make
sure this is a positive number.
3. Display the grid.

(a) Display the grid with the default settings (Figure 3.2).
(b) Use the middle mouse button to zoom in on the image so you can see the mesh
near the airfoil (Figure 3.3).
Quadrilateral cells were used for this simple geometry because they can be
stretched easily to account for different flow gradient magnitudes in different
directions. In the present case, the gradients normal to the airfoil wall are
much greater than those tangent to the airfoil, except near the leading and
trailing edges and in the vicinity of the shock expected on the upper surface.
Consequently, the cells nearest the surface have very high aspect ratios. For

Grid
Figure 3.2: The Grid Around the Airfoil
Grid
Figure 3.3: The Grid After Zooming In on the Airfoil
3-4
geometries that are more difficult to mesh, it may be easier to create a hybrid
mesh comprised of quadrilateral and triangular cells.
A parabola was chosen to represent the far-field boundary because it has no
discontinuities in slope, enabling the construction of a smooth mesh in the
interior of the domain.
Extra: You can use the right mouse button to check which zone number cor-
responds to each boundary. If you click the right mouse button on one
of the boundaries in the graphics window, its zone number, name, and
type will be printed in the FLUENT console window. This feature is espe-
cially useful when you have several zones of the same type and you want
to distinguish between them quickly.
4. Reorder the mesh.

Grid −→ Reorder −→Domain
This is done to reduce the bandwidth of the cell neighbor number and to speed up
the computations. This is especially important for large cases involving 1 million
or more cells.

Step 2: Models
1. Select the Coupled, Implicit solver.
The coupled solver is recommended when dealing with applications involving high-
speed aerodynamics with shocks. The implicit solver will generally converge much
faster than the explicit solver, but will use more memory. For this 2D case, memory
is not an issue.
3-6
2. Turn on the Spalart-Allmaras turbulence model.

(a) Select the Spalart-Allmaras model and retain the default options and constants.
The Spalart-Allmaras model is a relatively simple one-equation model that solves a mod-
eled transport equation for the kinematic eddy (turbulent) viscosity. This embodies a
relatively new class of one-equation models in which it is not necessary to calculate a
length scale related to the local shear layer thickness. The Spalart-Allmaras model was
designed specifically for aerospace applications involving wall-bounded flows and has been
shown to give good results for boundary layers subjected to adverse pressure gradients.

Step 3: Materials
The default Fluid Material is air, which is the working fluid in this problem. The default
settings need to be modified to account for compressibility and variations of the thermo-
physical properties with temperature.
1. Select ideal-gas in the Density drop-down list.
2. Select sutherland in the drop-down list for Viscosity.

This will open the Sutherland Law panel.
3-8
(a) Click OK to accept the default Three Coefficient Method and parameters.
The Sutherland law for viscosity is well suited for high-speed compressible flows.
3. Click Change/Create in the Materials panel to save these settings, and then close
the panel.
Note: While Density and Viscosity have been made temperature-dependent, Cp and Ther-
mal Conductivity have been left constant. For high-speed compressible flows, thermal
dependency of the physical properties is generally recommended. For simplicity,
Thermal Conductivity and Cp are assumed to be constant in this tutorial.

Step 4: Operating Conditions

Set the operating pressure to 0 Pa.
Define −→Operating Conditions...
When using the coupled solver for flows with Mach numbers greater than 0.1, an operating
pressure of zero is recommended.
See Section 8.13 of the User’s Guide for more information on how to set the operating
pressure.
3-10

Set the boundary conditions for pressure-far-field-1 as shown in the panel.
For external flows, you should choose a viscosity ratio between 1 and 10.
Note: The X- and Y-Component of Flow Direction are set as above because of the 4◦ angle
of attack: cos 4◦ ≈ 0.997564 and sin 4◦ ≈ 0.069756.

Step 6: Solution
1. Set the solution controls.
(a) Set the Under-Relaxation Factor for Modified Turbulent Viscosity to 0.9.
Larger (i.e., closer to 1) under-relaxation factors will generally result in faster
convergence. However, instability can arise that may need to be eliminated by
decreasing the under-relaxation factors.
(b) Under Solver Parameters, set the Courant Number to 5.
(c) Under Discretization, select Second Order Upwind for Modified Turbulent Vis-
cosity.
The second-order scheme will resolve the boundary layer and shock more ac-
curately than the first-order scheme.
3-12
2. Turn on residual plotting during the calculation.

(a) Select pressure-far-field-1 in the Compute From drop-down list.

(b) Click Init to initialize the solution.
To monitor the convergence of the solution, you are going to enable the plotting of
the drag, lift, and moment coefficients. You will need to iterate until all of these
forces have converged in order to be certain that the overall solution has converged.
For the first few iterations of the calculation, when the solution is fluctuating, the
values of these coefficients will behave erratically. This can cause the scale of the
y axis for the plot to be set too wide, and this will make variations in the value of
the coefficients less evident. To avoid this problem, you will have FLUENT perform
a small number of iterations, and then you will set up the monitors.
Since the drag, lift, and moment coefficients are global variables, indicating certain
overall conditions, they may converge while conditions at specific points are still
varying from one iteration to the next. To monitor this, you will create a point
monitor at a point where there is likely to be significant variation, just upstream
of the shock wave, and monitor the value of the skin friction coefficient. A small
number of iterations will be sufficient to roughly determine the location of the shock.
After setting up the monitors, you will continue the calculation.
4. Request 100 iterations.

This will be sufficient to see where the shock wave is, and the fluctuations of the
solution will have diminished significantly.
5. Increase the Courant number.
Under Solver Parameters, set the Courant Number to 20.
The solution will generally converge faster for larger Courant numbers, unless the
integration scheme becomes unstable. Since you have performed some initial iter-
ations, and the solution is stable, you can try increasing the Courant number to
speed up convergence. If the residuals increase without bound, or you get a floating
point exception, you will need to decrease the Courant number, read in the previous
data file, and try again.
6. Turn on monitors for lift, drag, and moment coefficients.
Solve −→ Monitors −→Force...
(a) In the drop-down list under Coefficient, select Drag.

(b) Select wall-bottom and wall-top in the Wall Zones list.
(c) Set Plot Window to 1.
(d) Under Force Vector, enter 0.9976 for X and 0.06976 for Y.
These magnitudes ensure that the drag and lift coefficients are calculated nor-
mal and parallel to the flow, which is 4◦ off of the global coordinates.
(e) Select Plot under Options to enable plotting of the drag coefficient.
(f) Select Write under Options to save the monitor history to a file, and specify
cd-history as the file name.
If you do not select the Write option, the history information will be lost when
you exit FLUENT.
3-14
(g) Click Apply.

(h) Repeat the above steps for Lift, using values of 0.06976 for X and 0.9976 for
Y under Force Vector and set Plot Window to 2.
(i) Repeat the above steps for Moment, using values of 0.25 m for X and 0 m for
Y under Moment Center and set Plot Window to 3.
7. Set the reference values that are used to compute the lift, drag, and moment coef-
ficients.
The reference values are used to non-dimensionalize the forces and moments acting
on the airfoil. The dimensionless forces and moments are the lift, drag, and moment
coefficients.
Report −→Reference Values...
(a) In the Compute From drop-down list, select pressure-far-field-1.
FLUENT will update the Reference Values based on the boundary conditions at
the far-field boundary.

8. Define a monitor for tracking the skin friction coefficient value just upstream of the
shock wave.
(a) Display filled contours of pressure overlaid with the grid.
i. Turn on Filled.
ii. Select Draw Grid.
This will open the Grid Display panel.
iii. Close the Grid Display panel, since there are no changes to be made here.
iv. Click Display in the Contours panel.
v. Zoom in on the airfoil (Figure 3.4).
1.55e+05
1.49e+05
1.44e+05
1.39e+05
1.33e+05
1.28e+05
1.23e+05
1.18e+05
1.12e+05
1.07e+05
1.02e+05
9.65e+04
9.12e+04
8.59e+04
8.06e+04
7.53e+04
7.00e+04
6.48e+04
5.95e+04
5.42e+04
4.89e+04

FLUENT 6.2 (2d, coupled imp, S-A)
Figure 3.4: Pressure Contours After 100 Iterations
The shock wave is clearly visible on the upper surface of the airfoil, where
the pressure first jumps to a higher value.
vi. Zoom in on the shock wave, until individual cells adjacent to the upper
surface (wall-top boundary) are visible (Figure 3.5).
3-16
1.55e+05
1.49e+05
1.44e+05
1.39e+05
1.33e+05
1.28e+05
1.23e+05
1.18e+05
1.12e+05
1.07e+05
1.02e+05
9.65e+04
9.12e+04
8.59e+04
8.06e+04
7.53e+04
7.00e+04
6.48e+04
5.95e+04
5.42e+04
4.89e+04

Figure 3.5: Magnified View of Pressure Contours Showing Wall-Adjacent Cells
The zoomed-in region contains cells just upstream of the shock wave that are
adjacent to the upper surface of the airfoil. In the following step, you will
create a point surface inside a wall-adjacent cell, to be used for the skin friction
coefficient monitor.
(b) Create a point surface just upstream of the shock wave.
Surface −→Point...
i. Under Coordinates, enter 0.53 for x0, and 0.051 for y0.
ii. Click on Create to create the point surface (point-4).

1.55e+05
1.49e+05
1.44e+05
1.39e+05
1.33e+05
1.28e+05
1.23e+05
1.18e+05
1.12e+05
1.07e+05
1.02e+05
9.65e+04
9.12e+04
8.59e+04
8.06e+04
7.53e+04
7.00e+04
6.48e+04
5.95e+04
5.42e+04
4.89e+04

Figure 3.6: Pressure Contours With Point Surface
Note: Here, you have entered the exact coordinates of the point surface so
that your convergence history will match the plots and description in this
tutorial. In general, however, you will not know the exact coordinates in
advance, so you will need to select the desired location in the graphics
window as follows:
i. Click Select Point With Mouse.
ii. Move the mouse to a point located anywhere inside one of the cells
adjacent to the upper surface (wall-top boundary), in the vicinity of
the shock wave. (See Figure 3.6.)
iii. Click the right mouse button.
iv. Click Create to create the point surface.
3-18
(c) Create a surface monitor for the point surface.

i. Increase Surface Monitors to 1.

ii. To the right of monitor-1, turn on the Plot and Write options and click
Define....
This will open the Define Surface Monitor panel.
iii. Select Wall Fluxes... and Skin Friction Coefficient under Report of.
iv. Select point-4 in the Surfaces list.
v. In the Report Type drop-down list, select Vertex Average.

vi. Increase the Plot Window to 4.

vii. Specify monitor-1.out as the File Name, and click OK in the Define Sur-
face Monitor panel.
viii. Click OK in the Surface Monitors panel.
9. Save the case file (airfoil.cas).

10. Continue the calculation by requesting 200 iterations.

Monitors
monitor-1
0.0025
0.0023
0.0020
0.0018
0.0015
Average 0.0013
of
0.0010
Surface
Vertex 0.0008
Values
0.0005
0.0003
0.0000
80 100 120 140 160 180 200 220 240
Iteration
Convergence history of Skin Friction Coefficient on point-4

Figure 3.7: Skin Friction Convergence History for the Initial Calculation
Note: After about 100 iterations, the residual criteria are satisfied and FLUENT
stops iterating. Since the skin friction monitor indicates that the skin friction
coefficient at point-4 has not converged (Figure 3.7), you will need to decrease
the convergence criterion for the modified turbulent viscosity and continue it-
erating.
3-20
11. Reduce the convergence criterion for the modified turbulent viscosity equation.
(a) Set the Convergence Criterion for nut to 1e-7 and click OK.
nut stands for νt . This is the residual for the modified turbulent viscosity that
the Spalart-Allmaras model solves for.
12. Continue the calculation for another 500 iterations.

After 500 additional iterations, the force monitors and the skin friction coefficient
monitor (Figures 3.8–3.11), indicate that the solution has reasonably converged.
13. Save the data file (airfoil.dat).


Monitors
monitor-1
0.0025
0.0023
0.0020
0.0018
0.0015
Average 0.0013
of
0.0010
Surface
Vertex 0.0008
Values
0.0005
0.0003
0.0000
100 200 300 400 500 600 700 800
Iteration
Convergence history of Skin Friction Coefficient on point-4

Figure 3.8: Skin Friction Coefficient History
0.0800
0.0750
0.0700
Cd 0.0650
0.0600
0.0550
0.0500
100 200 300 400 500 600 700 800
Iterations
Drag Convergence History

Figure 3.9: Drag Coefficient Convergence History
3-22
0.6000
0.5750
0.5500
0.5250
0.5000
0.4750
Cl
0.4500
0.4250
0.4000
0.3750
0.3500
0.3250
100 200 300 400 500 600 700 800
Iterations
Lift Convergence History

Figure 3.10: Lift Coefficient Convergence History
0.0700
0.0600
0.0500
0.0400
0.0300
Cm
0.0200
0.0100
0.0000
-0.0100
-0.0200
100 200 300 400 500 600 700 800
Iterations
Moment Convergence History About Z-Axis

Figure 3.11: Moment Coefficient Convergence History

1. Plot the y + distribution on the airfoil.
(a) Under Y Axis Function, select Turbulence... and Wall Yplus.

(b) In the Surfaces list, select wall-bottom and wall-top.
(c) Deselect Node Values and click Plot.
Wall Yplus is available only for cell values.
The values of y + are dependent on the resolution of the grid and the Reynolds
number of the flow, and are meaningful only in boundary layers. The value of y +
in the wall-adjacent cells dictates how wall shear stress is calculated. When you
use the Spalart-Allmaras model, you should check that y + of the wall-adjacent cells
is either very small (on the order of y + = 1), or approximately 30 or greater.
Otherwise, you will need to modify your grid.
The equation for y + is
y√
y+ = ρτw
µ
where y is the distance from the wall to the cell center, µ is the molecular viscosity,
ρ is the density of the air, and τw is the wall shear stress.
Figure 3.12 indicates that, except for a few small regions (notably at the shock and
the trailing edge), y + > 30 and for much of these regions it does not drop signifi-
cantly below 30. Therefore, you can conclude that the grid resolution is acceptable.
3-24
wall-bottom
wall-top
9.00e+01
8.00e+01
7.00e+01
6.00e+01
5.00e+01
Wall
Yplus 4.00e+01
3.00e+01
2.00e+01
1.00e+01
0.00e+00
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Position (m)
Wall Yplus
Figure 3.12: XY Plot of y + Distribution
2. Display filled contours of Mach number (Figure 3.13).

(a) Select Velocity... and Mach Number under Contours of.
(b) Turn off the Draw Grid option.
(c) Click Display.
Note the discontinuity, in this case a shock, on the upper surface at about
x/c ≈ 0.45.
3. Plot the pressure distribution on the airfoil (Figure 3.14).

(a) Under Y Axis Function, choose Pressure... and Pressure Coefficient from the
drop-down lists.
(b) Enable Node Values.
(c) Click Plot.
Notice the effect of the shock wave on the upper surface.
4. Plot the x component of wall shear stress on the airfoil surface (Figure 3.15).
(a) Under Y Axis Function, choose Wall Fluxes... and X-Wall Shear Stress from the
drop-down lists.

1.44e+00
1.37e+00
1.30e+00
1.22e+00
1.15e+00
1.08e+00
1.01e+00
9.39e-01
8.68e-01
7.96e-01
7.25e-01
6.53e-01
5.82e-01
5.10e-01
4.39e-01
3.68e-01
2.96e-01
2.25e-01
1.53e-01
8.17e-02
1.02e-02
Contours of Mach Number

Figure 3.13: Contour Plot of Mach Number
wall-bottom
wall-top
1.25e+00
1.00e+00
7.50e-01
5.00e-01
2.50e-01
Pressure 0.00e+00
Coefficient
-2.50e-01
-5.00e-01
-7.50e-01
-1.00e+00
-1.25e+00
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Position (m)
Pressure Coefficient
Figure 3.14: XY Plot of Pressure
3-26
(b) Disable Node Values.

(c) Click Plot.
wall-bottom
wall-top
2.25e+02
2.00e+02
1.75e+02
1.50e+02
1.25e+02
X-Wall 1.00e+02
Shear
7.50e+01
Stress
(pascal) 5.00e+01
2.50e+01
0.00e+00
-2.50e+01
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Position (m)
X-Wall Shear Stress

Figure 3.15: XY Plot of x Wall Shear Stress
The large, adverse pressure gradient induced by the shock causes the boundary layer
to separate. The point of separation is where the wall shear stress vanishes. Flow
reversal is indicated here by negative values of the x component of the wall shear
stress.
5. Display filled contours of the x component of velocity (Figure 3.16).

(a) Select Velocity... and X Velocity under Contours of.
(b) Click Display.
Note the flow reversal behind the shock.
6. Plot velocity vectors (Figure 3.17).

(a) Increase Scale to 15, and click Display.
Zooming in on the upper surface, behind the shock, will produce a display
similar to Figure 3.17. Flow reversal is clearly visible.

4.46e+02
4.20e+02
3.94e+02
3.68e+02
3.42e+02
3.16e+02
2.90e+02
2.64e+02
2.38e+02
2.12e+02
1.86e+02
1.60e+02
1.34e+02
1.08e+02
8.17e+01
5.56e+01
2.96e+01
3.57e+00
-2.25e+01
-4.85e+01
-7.45e+01
Contours of X Velocity (m/s)

Figure 3.16: Contour Plot of x Component of Velocity
4.48e+02
4.25e+02
4.03e+02
3.81e+02
3.58e+02
3.36e+02
3.14e+02
2.91e+02
2.69e+02
2.47e+02
2.24e+02
2.02e+02
1.80e+02
1.57e+02
1.35e+02
1.13e+02
9.05e+01
6.82e+01
4.59e+01
2.36e+01
1.27e+00

Figure 3.17: Plot of Velocity Vectors Near Upper Wall, Behind Shock
3-28
Summary
This tutorial demonstrated how to set up and solve an external aerodynamics problem us-
ing the Spalart-Allmaras turbulence model. It showed how to monitor convergence using
residual, force, and surface monitors, and demonstrated the use of several postprocessing
tools to examine the flow phenomena associated with a shock wave.

3-30
Tutorial 4. Modeling Unsteady Compressible Flow
Introduction
In this tutorial, FLUENT’s coupled implicit solver is used to predict the time-dependent
flow through a two-dimensional nozzle. As an initial condition for the transient problem,
a steady-state solution is generated to provide the initial values for the mass flow rate at
the nozzle exit.
• Calculate a steady-state solution (using the coupled implicit solver) as an initial

condition for a transient flow prediction
• Define an unsteady boundary condition using a user-defined function (UDF)
• Use dynamic mesh adaption for both steady-state and transient flows
• Calculate a transient solution using the second-order implicit unsteady formulation

and the coupled implicit solver
• Create an animation of the unsteady flow using FLUENT’s unsteady solution ani-
mation feature
Prerequisites
Problem Description
The geometry to be considered in this tutorial is shown in Figure 4.1. Flow through a
simple nozzle is simulated as a 2D planar model. The nozzle has an inlet height of 0.2 m,
and the nozzle contours have a sinusoidal shape that produces a 10% reduction in flow
area. Due to symmetry, only half of the nozzle is modeled.

Modeling Unsteady Compressible Flow
plane of symmetry
0.2 m p (t )
exit
p = 0.9 atm p = 0.7369 atm

inlet
exit
Setup and Solution

Preparation
1. Download unsteady_compressible.zip from the Fluent Inc. User Services Center
2. Unzip unsteady_compressible.zip.
nozzle.msh and pexit.c can be found in the /unsteady compressible folder cre-
ated after unzipping the file.
4-2
Step 1: Grid
1. Read in the mesh file nozzle.msh.
2. Check the grid.

Grid −→Check

To make the view more realistic, you will need to mirror it across the centerline.

4. Mirror the view across the centerline.

(a) Select symmetry under Mirror Planes.

(b) Click Apply.
The grid for the nozzle is shown in Figure 4.2.
Grid
Figure 4.2: 2D Nozzle Mesh Display
4-4
Step 2: Units
1. For convenience, define new units for pressure.
The pressure for this problem is specified in atm, which is not the default unit. You
will need to redefine the pressure unit as atm.
Define −→Units...
(a) Select pressure under Quantities, and atm under Units.

Hint: Use the scroll bar to access pressure, which is not initially visible in the
list.

Step 3: Models
1. Select the coupled implicit solver.
The coupled implicit solver is the solver of choice for compressible, transonic flows
without significant regions of low-speed flow. In cases with significant low-speed
flow regions, the segregated solver is preferred. Also, for transient cases with trav-
eling shocks, the coupled explicit solver with explicit time stepping may be the most
efficient.
Note: Initially, solve for the steady flow through the nozzle. Later, after obtaining
the steady flow as a starting point, this panel to enable an unsteady calculation.
2. Enable the energy equation.

4-6
3. Enable the Spalart-Allmaras turbulence model.

The Spalart-Allmaras model is a relatively simple one-equation model that solves

a modeled transport equation for the kinematic eddy (turbulent) viscosity. This
embodies a class of one-equation models in which it is not necessary to calculate a
length scale related to the local shear layer thickness. The Spalart-Allmaras model
was designed specifically for aerospace applications involving wall-bounded flows and
has been shown to give good results for boundary layers subjected to adverse pressure
gradients.

Step 4: Materials
1. Set the properties for air, the default fluid material.
(a) Select the ideal-gas law to compute Density.

Note: FLUENT will automatically enable solution of the energy equation when
the ideal gas law is used. You do not need to visit the Energy panel to turn
it on.
(b) Retain the default values for all other properties.
Do not forget to click the Change/Create button to save your change.

!
4-8

1. Set the operating pressure to 0 atm.
Here, the operating pressure is set to zero and boundary condition inputs for pressure
will be defined in terms of absolute pressures. Boundary condition inputs should
always be relative to the value used for operating pressure.


1. Set the conditions for the nozzle inlet (inlet).
(a) Set the Gauge Total Pressure to 0.9 atm.

(b) Set the Supersonic/Initial Gauge Pressure to 0.7369 atm.
The inlet static pressure estimate is the mean pressure at the nozzle exit. This
value will be used during the solution initialization phase to provide a guess
for the nozzle velocity.
(c) In the Turbulence Specification Method drop-down list, select Turbulent Viscosity
Ratio.
(d) Set the Turbulent Viscosity Ratio to 1.
For low to moderate inlet turbulence, a viscosity ratio of 1 is recommended.
4-10
2. Set the conditions for the nozzle exit (outlet).
(a) Set the Gauge Pressure to 0.7369 atm.

(b) In the Turbulence Specification Method drop-down list, select Turbulent Viscosity
Ratio.
(c) Accept the default value of 10 for Backflow Turbulent Viscosity Ratio.
If substantial backflow occurs at the outlet, you may need to adjust the backflow
values to levels close to the actual exit conditions.

Step 7: Solution: Steady Flow

(a) Select inlet in the Compute From drop-down list.

(b) Click Init, and Close the panel.
4-12

(a) Under Discretization, select Second Order Upwind for Modified Turbulent Vis-
cosity.
Second-order discretization provides optimum accuracy.
We will now activate dynamic adaption. Our purpose here is to have the solver
periodically refine the mesh in the vicinity of the shocks as the iterations progress.
We identify the shocks by looking for large gradients of the pressure.

3. Perform gradient adaption to refine the mesh.

Adapt −→Gradient
(a) Under Method, select Gradient.

The mesh adaption criterion can either be the gradient or the curvature (second
gradient). Because strong shocks occur inside the nozzle, the gradient is used
as the adaption criterion.
(b) Under Gradients Of, make sure that Pressure... and Static Pressure are selected.
(c) Under Normalization, select Scale.
Mesh adaption can be controlled by the raw (or standard) value of the gradient,
the scaled value (by its average in the domain), or the normalized value (by its
maximum in the domain). For dynamic mesh adaption, it is recommended to
use either the scaled or normalized value because the raw values will probably
change strongly during the computation, which would necessitate a readjust-
ment of the coarsen and refine thresholds. In this case, the scaled gradient is
used.
(d) Set the Coarsen Threshold to 0.3.
(e) Set Refine Threshold to 0.7.
As the refined regions of the mesh get larger, the coarsen and refine thresholds
should get smaller. A coarsen threshold of 0.3 and a refine threshold of 0.7
result in a “medium” to “strong” mesh refinement in combination with the
scaled gradient.
4-14
(f) Turn on the Dynamic option under Dynamic and set the Interval to 100.
For steady-state flows, it is sufficient to only seldomly adapt the mesh—in
this case an interval of 100 iterations is chosen. For time-dependent flows, a
considerably smaller interval must be used.
(g) Click Apply to store the information.
(h) Click on Controls... to modify the adaption controls.
i. Make sure that fluid is selected under Zones.

ii. Set the Max # of Cells to 20000.
To restrict the mesh adaption, the maximum number of cells can be lim-
ited. If this limit is violated during the adaption, the coarsen and refine
thresholds are adjusted to respect the maximum number of cells. Addi-
tional restrictions can be placed on the minimum cell volume, minimum
number of cells, and maximum level of refinement.
iii. Click OK.


(a) Under Options, select Plot.

(b) Click OK.
5. Enable the plotting of mass flow rate at the flow exit.
4-16

(b) Turn on the Plot and Write options for monitor-1.
Note: When the Write option is selected in the Surface Monitors panel, the
mass flow rate history will be written to a file. If you do not select the
write option, the history information will be lost when you exit FLUENT.
(c) Click on Define... to specify the surface monitor parameters in the Define
Surface Monitor panel.
i. Select Mass Flow Rate in the Report Type drop-down list.

ii. Select outlet in the Surfaces list.
iii. In the File Name field, enter the name noz ss.out.
iv. Click on OK to define the monitor.
(d) Click on OK in the Surface Monitors panel to enable the monitor.
6. Save the case file (noz ss.cas).



The mass flow rate history is shown in Figure 4.3. It shows that the solution is
converged after around 1800 iterations.
-13.5000
-14.0000
-14.5000
-15.0000
Mass
Flow -15.5000
Rate
(kg/s)
-16.0000
-16.5000
-17.0000
0 200 400 600 800 1000 1200 1400 1600 1800 2000
Iteration
Convergence history of Mass Flow Rate on outlet (in SI units)

Figure 4.3: Mass Flow Rate History
8. Save the data file (noz ss.dat).

4-18
9. Check the mass flux balance.

Report −→Fluxes...
Although the mass flow rate history indicates that the solution is con-
! verged, you should also check the mass fluxes through the domain to ensure
that mass is being conserved.
(a) Keep the default Mass Flow Rate option.

(b) Select inlet and outlet in the Boundaries list.
(c) Click Compute.
The net mass imbalance should be a small fraction (say, 0.5%) of the total
!
flux through the system. If a significant imbalance occurs, you should
decrease your residual tolerances by at least an order of magnitude and
continue iterating.

10. Display the steady-flow velocity vectors (Figure 4.4).

(a) Change the Scale to 10.

(b) In the Surfaces list, select all of the surfaces.
(c) Click Display.
The steady flow prediction shows the expected form, with peak velocity of about 335
m/s through the nozzle.
4-20
3.45e+02
3.27e+02
3.10e+02
2.93e+02
2.76e+02
2.59e+02
2.41e+02
2.24e+02
2.07e+02
1.90e+02
1.73e+02
1.55e+02
1.38e+02
1.21e+02
1.04e+02
8.64e+01
6.92e+01
5.20e+01
3.48e+01
1.76e+01
3.72e-01

Figure 4.4: Velocity Vectors (Steady Flow)

11. Display the steady flow contours of static pressure (Figure 4.5).

(b) Click Display.
The steady flow prediction shows the expected pressure distribution, with low pres-
sure near the nozzle throat.
4-22
7.84e-01
7.66e-01
7.47e-01
7.28e-01
7.09e-01
6.90e-01
6.71e-01
6.53e-01
6.34e-01
6.15e-01
5.96e-01
5.77e-01
5.59e-01
5.40e-01
5.21e-01
5.02e-01
4.83e-01
4.65e-01
4.46e-01
4.27e-01
4.08e-01
Contours of Static Pressure (atm)

Figure 4.5: Contours of Static Pressure (Steady Flow)

Step 8: Enable Time Dependence and Set Unsteady Conditions

In this step you will define a transient flow by specifying an unsteady pressure condition
for the nozzle.
1. Enable a time-dependent flow calculation.

(a) Under Time, select Unsteady.

(b) Under Unsteady Formulation, select 2nd-Order Implicit.
Implicit (dual) time-stepping allows you to set the physical time step used for the
transient flow prediction (while FLUENT continues to determine the time step used
for inner iterations based on a Courant condition). Here, second-order implicit
time-stepping is enabled: this provides higher accuracy in time than the first-order
option.
4-24
2. Define the unsteady condition for the nozzle exit (outlet).

The pressure at the outlet is defined as a wave-shaped profile, and is described by
the following equation:
pexit (t) = 0.12 sin(ωt) + pexit (4.1)
where
ω = circular frequency of unsteady pressure (rad/s)
pexit = mean exit pressure (atm)
In this case, ω = 2200 rad/s, and pexit = 0.7369 atm.
A user-defined function (pexit.c) has been written to define the equation (Equa-
tion 4.1) required for the pressure profile.
Note: To input the value of Equation 4.1 in the correct units, the function pexit.c
has been multiplied by a factor of 101325 to convert from the chosen pressure
unit (atm) to the SI unit required by FLUENT (Pa). This will not affect the
displayed results.
See the separate UDF Manual for details about user-defined functions..
(a) Read in the user-defined function.
Define −→ User-Defined −→ Functions −→Interpreted...
i. Enter pexit.c as the Source File Name.

ii. Click Interpret.
The user-defined function has already been defined, but it needs to be com-
piled within FLUENT before it can be used in the solver.
iii. Close the Interpreted UDFs panel.

(b) Set the unsteady boundary conditions at the exit.
i. Select udf unsteady pressure (the user-defined function) in the Gauge Pres-
sure drop-down list.
3. Update the gradient adaption parameters for the transient case.

Adapt −→Gradient
(a) Reset the Coarsen Threshold to 0.3.
(b) Reset the Refine Threshold to 0.7.
The refine and coarsen thresholds have been changed during the steady-state
computation to meet the limit of 20000 cells. Therefore, you need to reset
these parameters to their original values.
(c) Under Dynamic, set the Interval to 1.
For the transient case, the mesh adaption will be done every time step.
(d) Click Apply to store the values.
(e) Click on Controls... to modify the adaption controls.
i. In the Grid Adaption Controls panel, set the Min # of Cells to 8000.
ii. Set the Max # of Cells to 30000 and click OK.
The maximum number of cells is increased to try to avoid readjustment
of the coarsen and refine thresholds. Additionally, the minimum number
of cells has been limited to 8000 because it is not desired to have a coarse
mesh during the computation (the current mesh has about 10000 cells).
4-26
Step 9: Solution: Unsteady Flow

1. Set the time step parameters.
The selection of the time step is critical for accurate time-dependent flow predic-
tions. Using a time step of 2.85596 × 10−5 seconds, 100 time steps are required for
one pressure cycle. The pressure cycle begins and ends with the initial pressure at
the nozzle exit.
(a) Set the Time Step Size to 2.85596e-5 s.

(b) Set the Number of Time Steps to 600.
(c) Set the Max Iterations per Time Step to 30.
(d) Click Apply.

2. Modify the plotting of the mass flow rate at the nozzle exit.
Because each time step requires 30 iterations, a smoother plot will be generated by
plotting at every time step.
(a) For monitor-1, select Time Step in the drop-down list under Every.
(b) Click Define... to modify the surface monitor parameters.
i. In the Define Surface Monitor panel, change the File Name to noz uns.out.
ii. In the X Axis drop-down list, select Time Step.
iii. Click OK.
(c) Click OK in the Surface Monitors panel.
3. Save the transient solution case file (noz uns.cas).

4-28
4. Start the transient calculation.

Calculation of 600 time steps will require significant CPU resources (about
!
three hours on a single CPU with a 2.6 GHz clock speed). Instead of
calculating, you can read the data file saved after the iterations have been
completed:
noz uns.dat
(The data file is available in the same directory where you found the mesh
and UDF files.)
By requesting 600 time steps, you are asking FLUENT to compute six pressure
cycles. The mass flow rate history is shown in Figure 4.6.
Monitors
monitor-1
-4.0000
-6.0000
-8.0000
-10.0000
Mass -12.0000
Flow
Rate -14.0000
(kg/s)
-16.0000
-18.0000
-20.0000
0 100 200 300 400 500 600
Time Step
Convergence history of Mass Flow Rate on outlet (Time=1.7136e-02)

FLUENT 6.2 (2d, coupled imp, S-A, unsteady)
Figure 4.6: Mass Flow Rate History (Unsteady Flow)
5. Save the transient solution data file (noz uns.dat).


Step 10: Saving and Postprocessing Time-Dependent Data Sets

The solution has reached a time-periodic state. To study how the flow changes within a
single pressure cycle, you will now continue the solution for 100 more time steps. You
will use FLUENT’s solution animation feature to save contour plots of pressure and Mach
number at each time step, and the autosave feature to save case and data files every 10
time steps. After the calculation is complete, you will use the solution animation playback
feature to view the animated pressure and Mach number plots over time.
1. Request saving of case and data files every 10 time steps.

File −→ Write −→Autosave...
(a) Set the Autosave Case File Frequency and Autosave Data File Frequency to 10.
(b) In the Filename field, enter noz anim.
(c) Click OK.
If you have constraints on the disc space, you can restrict the number of files
saved by FLUENT using the Maximum Number of Files at Any Instance field.
After saving the specified number of files, FLUENT will overwrite the earliest
existing file. The default value of zero will save all the files.
When FLUENT saves a file, it will append the time step value to the file
name prefix (noz anim). The standard extensions (.cas and .dat) will also
be appended. This will yield file names of the form noz anim0640.cas and
noz anim0640.dat, where 0640 is the time step number.
Optionally, you can add the extension .gz to the end of the file name (e.g.,
noz anim.gz), which will instruct FLUENT to save the case and data files in
compressed format, yielding file names of the form noz anim0640.cas.gz.
4-30
2. Create animation sequences for the nozzle pressure and Mach number contour plots.
Solve −→ Animate −→Define...
(a) Increase the number of Animation Sequences to 2.

(b) Under Name, enter pressure for the first sequence and mach-number for the
second sequence.
(c) In the When drop-down lists, select Time Step.
With the default value of 1 for Every, this instructs FLUENT to update the
animation sequence at every time step.

3. Define the animation sequence for pressure.

(a) Click Define... on the line for pressure to set the parameters for the pressure
sequence.
The Animation Sequence panel will open.
(b) Under Storage Type, select In Memory.

The Memory option is acceptable for a small 2D case such as this. For larger
2D or 3D cases, saving animation files with either the Metafile or PPM Image
option is preferable to avoid using too much of your machine’s memory.
(c) Increase the Window number to 2 and click Set.
Graphics window number 2 will open.
(d) Under Display Type, select Contours.
The Contours panel will open.
4-32
i. In the Contours panel, keep the default selections of Pressure... and Static
Pressure.
ii. Make sure that Filled is selected under Options, and deselect Auto Range.
iii. Enter 0.25 under Min and 1.25 under Max.
This will set a fixed range for the contour plot and subsequent animation.
iv. In the Surfaces list, select all of the surfaces.
v. Click Display.
Figure 4.7 shows the contours of static pressure in the nozzle after 600 time
steps.
(e) Click OK in the Animation Sequence panel.

1.25e+00
1.20e+00
1.15e+00
1.10e+00
1.05e+00
1.00e+00
9.50e-01
9.00e-01
8.50e-01
8.00e-01
7.50e-01
7.00e-01
6.50e-01
6.00e-01
5.50e-01
5.00e-01
4.50e-01
4.00e-01
3.50e-01
3.00e-01
2.50e-01
Contours of Static Pressure (atm) (Time=1.7136e-02)

Figure 4.7: Pressure Contours at t = 0.01714 s
4-34
4. Define the animation sequence for Mach number.

(a) In the Solution Animation panel, click Define... on the line for mach-number to
set the parameters for the Mach number sequence.
(b) Under Storage Type in the Animation Sequence panel, select In Memory.
(c) Increase the Window number to 3 and click Set.
(d) Under Display Type, select Contours.
i. In the Contours panel, select Velocity... and Mach Number.
ii. Make sure that Filled is selected under Options, and deselect Auto Range.
iii. Enter 0.00 under Min and 1.30 under Max.
iv. In the Surfaces list, make sure that all of the surfaces are selected.
v. Click Display.
Figure 4.8 shows the Mach number contours in the nozzle after 600 time steps.
1.30e+00
1.23e+00
1.17e+00
1.11e+00
1.04e+00
9.75e-01
9.10e-01
8.45e-01
7.80e-01
7.15e-01
6.50e-01
5.85e-01
5.20e-01
4.55e-01
3.90e-01
3.25e-01
2.60e-01
1.95e-01
1.30e-01
6.50e-02
0.00e+00
Contours of Mach Number (Time=1.7136e-02)

Figure 4.8: Mach Number Contours at t = 0.01714 s
(e) Click OK in the Animation Sequence panel.

(f) Click OK in the Solution Animation panel.

5. Continue the calculation by requesting 100 time steps.

Requesting 100 time steps will march the solution through an additional 0.0028
seconds, or roughly one pressure cycle. This will take about 30 minutes. With the
autosave and animation features active (as defined above), the case and data files
will be saved approximately every 0.00028 seconds; animation files will be saved
every 0.000028 seconds.
When the calculation finishes, you will have ten pairs of case and data files and
there will be 100 pairs of contour plots stored in memory. In the next few steps,
you will play back the animation sequences and examine the results at several time
steps after reading in pairs of newly-saved case and data files.
4-36
6. Change the display options to include double buffering.

Double buffering will allow for a smoother transition between the frames of the
animations.
Display −→Options...
(a) Under Rendering, select Double Buffering.

(b) Click Apply.

7. Play back the animation of the pressure contours.

Solve −→ Animate −→Playback...
(a) Under Sequences, select pressure.

The playback control buttons now become active.
(b) Keep the default settings in the rest of the panel and click the play button
(the second from the right in the group of buttons under Playback).
Examples of pressure contours at t = 0.01799 s (630th time step) and t = 0.0191 s
(670th time step) are shown in Figures 4.9 and 4.10.
8. Repeat steps 6 and 7, selecting the appropriate active window and sequence name
for the Mach number contours.
Examples of Mach number contours at t = 0.01799 s and t = 0.0191 s are shown
in Figures 4.11 and 4.12.
4-38
1.25e+00
1.20e+00
1.15e+00
1.10e+00
1.05e+00
1.00e+00
9.50e-01
9.00e-01
8.50e-01
8.00e-01
7.50e-01
7.00e-01
6.50e-01
6.00e-01
5.50e-01
5.00e-01
4.50e-01
4.00e-01
3.50e-01
3.00e-01
2.50e-01

1.25e+00
1.20e+00
1.15e+00
1.10e+00
1.05e+00
1.00e+00
9.50e-01
9.00e-01
8.50e-01
8.00e-01
7.50e-01
7.00e-01
6.50e-01
6.00e-01
5.50e-01
5.00e-01
4.50e-01
4.00e-01
3.50e-01
3.00e-01
2.50e-01


1.30e+00
1.23e+00
1.17e+00
1.11e+00
1.04e+00
9.75e-01
9.10e-01
8.45e-01
7.80e-01
7.15e-01
6.50e-01
5.85e-01
5.20e-01
4.55e-01
3.90e-01
3.25e-01
2.60e-01
1.95e-01
1.30e-01
6.50e-02
0.00e+00

1.30e+00
1.23e+00
1.17e+00
1.11e+00
1.04e+00
9.75e-01
9.10e-01
8.45e-01
7.80e-01
7.15e-01
6.50e-01
5.85e-01
5.20e-01
4.55e-01
3.90e-01
3.25e-01
2.60e-01
1.95e-01
1.30e-01
6.50e-02
0.00e+00

4-40
Extra: FLUENT gives you the option of exporting an animation as an MPEG file
or as a series of files in any of the hardcopy formats available in the Graphics
Hardcopy panel (including TIFF and PostScript).
To save an MPEG file, select MPEG from the Write/Record Format drop-down
list in the Playback panel and then click the Write button. The MPEG file will
be saved in your working directory. You can view the MPEG movie using an
MPEG player (e.g., Windows Media Player or another MPEG movie player).
To save a series of TIFF, PostScript, or other hardcopy files, select Hardcopy
Frames in the Write/Record Format drop-down list in the Playback panel. Click
on the Hardcopy Options... button to open the Graphics Hardcopy panel and set
the appropriate parameters for saving the hardcopy files. Click Apply in the
Graphics Hardcopy panel to save your modified settings. In the Playback panel,
click the Write button. FLUENT will replay the animation, saving each frame
to a separate file in your working directory.
If you want to view the solution animation in a later FLUENT session, you
can select Animation Frames as the Write/Record Format and click Write.
Since the solution animation was stored in memory, it will be lost if you
! exit FLUENT without saving it to one of the formats described above. Note
that only the animation-frame format can be read back into the Playback
panel for display in a later FLUENT session.
9. Display velocity vectors after 60 time steps (Figure 4.13).

(a) Read case and data files for the 660th time step (noz anim0660.cas and
noz anim0660.dat) into FLUENT.
File −→ Read −→Case & Data...
(b) Plot vectors at t = 0.01885 s.

i. Change the Scale to 10.

ii. Click Display.
The unsteady flow prediction shows the expected form, with peak velocity of
about 241 m/s through the nozzle at t = 0.01885 seconds.
10. Repeat step 9 using case and data files saved for other time steps of interest.
4-42
2.42e+02
2.30e+02
2.18e+02
2.06e+02
1.94e+02
1.82e+02
1.70e+02
1.58e+02
1.46e+02
1.34e+02
1.22e+02
1.10e+02
9.75e+01
8.54e+01
7.34e+01
6.13e+01
4.92e+01
3.71e+01
2.50e+01
1.30e+01
8.97e-01
Velocity Vectors Colored By Velocity Magnitude (m/s) (Time=1.8849e-02)

Figure 4.13: Velocity Vectors at t = 0.01885 s
Summary
In this tutorial, you modeled the transient flow of air through a nozzle. You learned how
to generate a steady-state solution as an initial condition for the unsteady case, and how
to set solution parameters for implicit time-stepping.
You also learned how to manage the file saving and graphical postprocessing for time-
dependent flows, using file autosaving to automatically save solution information as the
transient calculation proceeds.
Finally, you learned how to use FLUENT’s solution animation tool to create animations
of transient data, and how to view the animations using the playback feature.

4-44
Tutorial 5. Modeling Radiation and Natural Convection
Introduction
In this tutorial, combined radiation and natural convection are solved in a two-dimensional
square box on a mesh consisting of quadrilateral elements.
• Use the radiation models in FLUENT (Rosseland, P-1, DTRM, discrete ordinates
(DO), and surface-to-surface (S2S)) and understand their ranges of application
• Use the Boussinesq model for density
• Set the boundary conditions for a heat transfer problem involving natural convec-
tion and radiation
• Separate a single wall zone into multiple wall zones
• Change the properties of an existing fluid material
• Display velocity vectors and contours of stream function and temperature for flow
visualization
Prerequisites
Problem Description
The problem to be considered is shown schematically in Figure 5.1. A square box of side
L has a hot right wall at T = 2000 K, a cold left wall at T = 1000 K, and adiabatic
top and bottom walls. Gravity points downwards. A buoyant flow develops because of
thermally-induced density gradients. The medium contained in the box is assumed to be
absorbing and emitting, so that the radiant exchange between the walls is attenuated by
absorption and augmented by emission in the medium. All walls are black. The objective
is to compute the flow and temperature patterns in the box, as well as the wall heat flux,

Modeling Radiation and Natural Convection
using the radiation models available in FLUENT, and to compare their performance for
different values of the optical thickness aL.
The working fluid has a Prandtl number of approximately 0.71, and the Rayleigh number
based on L is 5 × 105 . This means the flow is inherently laminar. The Boussinesq
assumption is used to model buoyancy. The Planck number k/(4σLT03 ) is 0.02, and
measures the relative importance of conduction to radiation; here T0 = (Th + Tc )/2.
Three values for the optical thickness are considered: aL = 0, aL = 0.2, and aL = 5.
Note that the values of physical properties and operating conditions (e.g., gravitational
acceleration) have been adjusted to yield the desired Prandtl, Rayleigh, and Planck
numbers.
Adiabatic
ρ = 1000 kg/m3
4
cp= 1.1030x10 J/kgK
k = 15.309 W/mK
-3
µ = 10 kg/ms
Tc= 1000 K
T = 2000 K
-5
h
g ➢ β = 10 1/K
➢
-5
g = -6.96 x 10 m/s2
a = 0, 0.2, 5 1/m
y L=1m
5
x Ra = 5 x 10
Pr = 0.71
Pl = 0.02
L τ = 0.2, 5
Setup and Solution

Preparation
1. Download radiation_natural_convection.zip from the Fluent Inc. User Ser-
vices Center or copy it from the FLUENT documentation CD to your working
directory (as described in Tutorial 1).
2. Unzip radiation_natural_convection.zip.
rad.msh can be found in the /radiation natural convection folder created after
unzipping the file.
5-2
Step 1: Grid
1. Read the mesh file rad.msh.
As the mesh is read in, messages will appear in the console window reporting the
progress of the reading. The mesh size will be reported as 2500 cells.
2. Check the grid.

Grid −→Check
FLUENT performs various checks on the mesh and reports the progress in the console
window. Pay particular attention to the minimum volume. Make sure this is a
positive number.

Note: All the walls are currently contained in a single wall zone, wall-4. You will
need to separate them out into four different walls so that you can specify
different boundary conditions for each wall.

Grid
Figure 5.2: Graphics Display of Grid
4. Separate the single wall zone into four wall zones.

Grid −→ Separate −→Faces...
(a) Select the Angle separation method (the default) under Options.
(b) Select wall-4 in the Zones list.
(c) Specify 89◦ as the significant Angle.
(d) Click on the Separate button.
Faces with normal vectors that differ by more than 89◦ will be placed in separate
zones. Since the four wall zones are perpendicular (angle = 90◦ ), wall-4 will be
separated into four zones.
5-4
5. Display the grid again.

(a) Select all Surfaces and click on Display.
Notice that you now have four different wall zones instead of only one.
Extra: You can use the right mouse button to check which wall zone number
corresponds to each wall boundary. If you click the right mouse button
on one of the boundaries in the graphics window, its zone number, name,
and type will be printed in the FLUENT console window. This feature is
especially useful when you have several zones of the same type and you
want to distinguish between them quickly. In some cases, you may want
to disable the display of the interior grid so as to more accurately select
the boundaries for identification.

Step 2: Models
As discussed earlier, in this tutorial you will enable each radiation model in turn, obtain
a solution, and postprocess the results. You will start with the Rosseland model, then use
the P-1 model, the discrete transfer radiation model (DTRM), and the discrete ordinates
(DO) model. At the end of the tutorial, you will use the surface-to-surface (S2S) model.

5-6
2. Turn on the Rosseland radiation model.

Define −→ Models −→Radiation...
When you click OK in the Radiation Model panel, FLUENT will present an Infor-
mation dialog box telling you that new material properties have been added for the
radiation model. You will be setting properties later, so you can simply click OK in
the dialog box to acknowledge this information.
Note: FLUENT will automatically enable the energy calculation when you enable
a radiation model, so you need not visit the Energy panel.
3. Add the effect of gravity on the model.

(a) Turn on Gravity.

The panel will expand to show additional inputs.
(b) Set the Gravitational Acceleration in the Y direction to -6.94e-5 m/s2 .
As mentioned earlier, the gravitational acceleration has been adjusted to yield
the correct dimensionless quantities (Prandtl, Rayleigh, and Planck numbers).
See Figure 5.1 and the associated comments.
(c) Set the Operating Temperature to 1000 K.
The operating temperature will be used by the Boussinesq model, which you
will enable in the next step.
5-8
Step 3: Materials
The default fluid material is air, which is the working fluid in this problem. However,
since you are working with a fictitious fluid whose properties have been adjusted to give
the desired values of the dimensionless parameters, you must change the default properties
for air. You will use an optical thickness aL of 0.2 for this calculation. (Since L = 1, the
absorption coefficient a will be set to 0.2.) Later in the tutorial, results for an optically
thick medium with aL = 5 and non-participating medium with aL = 0 are computed to
show how the different radiation models behave for different optical thicknesses.
1. Select boussinesq in the drop-down list next to Density, and then set the Density to
1000 kg/m3 .
For details about the Boussinesq model, see the User’s Guide.
2. Set the specific heat, Cp, to 1.103e4 J/kg-K.
3. Set the Thermal Conductivity to 15.309 W/m-K.

4. Set the Viscosity to 0.001 kg/m-s.
5. Set the Absorption Coefficient to 0.2 m−1 .

Hint: Use the scroll bar to access the properties that are not initially visible in the
panel.
6. Keep the default settings for the Scattering Coefficient and the Scattering Phase
Function, since there is no scattering in this problem.
7. Set the Thermal Expansion Coefficient (used by the Boussinesq model) to 1e-5 K−1 .
8. Click on Change/Create and close the Materials panel.
5-10

1. Set the boundary conditions for the bottom wall (wall-4.006).

Note: The bottom wall should be called wall-4.006, but to be sure that you have the
correct wall, use your right mouse button to click on the bottom wall in the
graphics window. When you do this, the corresponding zone will be selected
automatically in the Zone list in the Boundary Conditions panel. You can do
this when you set boundary conditions for the other walls as well, to be sure
that you are defining the correct conditions.
(a) Change the Zone Name to bottom.

(b) Retain the default thermal conditions (heat flux of 0) to specify an adiabatic
wall.
Note: The Rosseland model does not require you to set a wall emissivity. Later in
the tutorial, you will need to define the wall emissivity for the other radiation
models.
2. Set the boundary conditions for the left wall, wall-4.

(a) Change the Zone Name to left.
(b) Select Temperature under Thermal Conditions and set the Temperature to
1000 K.

3. Set the boundary conditions for the right wall, wall-4:007.

(a) Change the Zone Name to right.
(b) Select Temperature under Thermal Conditions and set the Temperature to
2000 K.
4. Set the boundary conditions for the top wall, wall-4:005.

(a) Change the Zone Name to top.
(b) Retain the default thermal conditions (heat flux of 0) to specify an adiabatic
wall.
5-12
Step 5: Solution for the Rosseland Model

1. Set the parameters that control the solution.
(a) Retain the default selected Equations (all of them) and Under-Relaxation Fac-
tors.
(b) Under Discretization, select PRESTO! for Pressure, and Second Order Upwind
for Momentum and Energy.

2. Initialize the flow field.

(a) Set the Temperature to 1500 K and click on Init.

5-14

(b) Click OK.
Note: There is no extra residual for the radiation heat transfer because the Rosse-
land model does not solve extra transport equations for radiation; instead, it
augments the thermal conductivity in the energy equation. When you use the
P-1 and DO radiation models, which both solve additional transport equations,
you will see additional residuals for radiation.
4. Save the case file (rad ross.cas).


The solution will converge in about 180 iterations.
6. Save the data file (rad ross.dat).


Step 6: Postprocessing for the Rosseland Model

5-16
2.10e-04
2.00e-04
1.89e-04
1.79e-04
1.68e-04
1.58e-04
1.47e-04
1.37e-04
1.26e-04
1.16e-04
1.05e-04
9.45e-05
8.40e-05
7.35e-05
6.30e-05
5.25e-05
4.20e-05
3.15e-05
2.10e-05
1.05e-05
3.26e-09

Figure 5.3: Velocity Vectors for the Rosseland Model
2. Display contours of stream function (Figure 5.4).


The recirculatory patterns observed are due to the natural convection in the box.
At a low optical thickness (0.2), radiation should not have a large influence on the
flow. The flow pattern is expected to be similar to that obtained with no radiation
(Figure 5.5). However, the Rosseland model predicts a flow pattern that is very
symmetric (Figure 5.4), and quite different from the pure natural convection case.
This discrepancy occurs because the Rosseland model is not appropriate for small
optical thickness.
6.98e-02
6.63e-02
6.28e-02
5.93e-02
5.58e-02
5.23e-02
4.88e-02
4.54e-02
4.19e-02
3.84e-02
3.49e-02
3.14e-02
2.79e-02
2.44e-02
2.09e-02
1.74e-02
1.40e-02
1.05e-02
6.98e-03
3.49e-03
0.00e+00
Contours of Stream Function (kg/s)

Figure 5.4: Contours of Stream Function for the Rosseland Model
Extra: If you want to compute the results without radiation yourself, turn off all
the radiation models in the Radiation Model panel, set the under-relaxation
factor for energy to 0.8, and calculate until convergence. (Remember to reset
the under-relaxation factor to 1 before continuing with the tutorial).
5-18
1.97e-02
1.87e-02
1.77e-02
1.67e-02
1.58e-02
1.48e-02
1.38e-02
1.28e-02
1.18e-02
1.08e-02
9.85e-03
8.86e-03
7.88e-03
6.89e-03
5.91e-03
4.92e-03
3.94e-03
2.95e-03
1.97e-03
9.85e-04
0.00e+00

Figure 5.5: Contours of Stream Function with No Radiation


2.00e+03
1.95e+03
1.90e+03
1.85e+03
1.80e+03
1.75e+03
1.70e+03
1.65e+03
1.60e+03
1.55e+03
1.50e+03
1.45e+03
1.40e+03
1.35e+03
1.30e+03
1.25e+03
1.20e+03
1.15e+03
1.10e+03
1.05e+03
1.00e+03

Figure 5.6: Contours of Temperature for the Rosseland Model
The Rosseland model predicts a temperature field (Figure 5.6) very different from
that obtained without radiation (Figure 5.7). For the low optical thickness in this
problem, the temperature field predicted by the Rosseland model is not physical.
4. Plot the y velocity along the horizontal centerline of the box.
(a) Create an isosurface at y = 0.5, the horizontal line through the center of the
box.
5-20
2.00e+03
1.95e+03
1.90e+03
1.85e+03
1.80e+03
1.75e+03
1.70e+03
1.65e+03
1.60e+03
1.55e+03
1.50e+03
1.45e+03
1.40e+03
1.35e+03
1.30e+03
1.25e+03
1.20e+03
1.15e+03
1.10e+03
1.05e+03
1.00e+03

Figure 5.7: Contours of Temperature with No Radiation
i. Select Grid... in the Surface of Constant drop-down list and select Y-

Coordinate from the list below.
ii. Click on Compute to see the extents of the domain.
iii. Set a value of 0.5 in the Iso-Values field, and change the New Surface
Name to y=0.5.
iv. Click on Create to create a surface at y = 0.5.

(b) Create an XY plot of y velocity on the isosurface.

i. Make sure that Node Values is turned on under Options.

By default, the Node Values option is turned on, and the values that have
been interpolated to the nodes are displayed. If you prefer to display the cell
values, turn the Node Values option off. Note that you will need to ensure
that the selected option for Node Values is used throughout the tutorial for
displaying and saving XY plots. This will enable you to correctly compare
the XY plots for different radiation models in a later step, as they will use
identical options.
ii. Check that the Plot Direction for X is 1, and the Plot Direction for Y is 0.
With a Plot Direction vector of (1,0), FLUENT will plot the selected vari-
able as a function of x. Since you are plotting the velocity profile on a
cross-section of constant y, the x direction is the one in which the velocity
varies.
iii. Select Velocity... and Y Velocity under Y Axis Function.
iv. Select y=0.5 in the Surfaces list.
v. Click on Plot.
The velocity profile reflects the rising plume at the hot right wall, and the
falling plume at the cold left wall. Compared to the case with no radiation,
5-22
y=0.5
2.50e-04
2.00e-04
1.50e-04
1.00e-04
5.00e-05
Y 0.00e+00
Velocity
-5.00e-05
(m/s)
-1.00e-04
-1.50e-04
-2.00e-04
-2.50e-04
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Position (m)
Y Velocity
Figure 5.8: XY Plot of Centerline y Velocity for the Rosseland Model
the profile predicted by the Rosseland model exhibits thicker wall layers.
As discussed before, the expected profile for aL = 0.2 is similar to the case
with no radiation.
(c) Save the plot data to a file.
i. Select the Write to File option, and click the Write... button.
ii. In the resulting Select File dialog box, specify rad ross.xy in the XY File
text entry box and click OK.

5. Compute the total wall heat flux on each lateral wall.

(a) Select Total Heat Transfer Rate under Options.

(b) Select right and left under Boundaries.
(c) Click the Compute button.
The total wall heat transfer rate is reported for the hot and cold walls as ap-
proximately 7.43 × 105 W. The sum of the heat fluxes on the lateral walls is a
negligible imbalance.
6. Save the case and data files (rad ross.cas and rad ross.dat).
Thus far in this tutorial, you have learned how to set up a natural convection problem
using the Rosseland model to compute radiation. You have also learned to postprocess the
results. You will now turn on the P-1 model and compare the results so computed with
those of the Rosseland model.
5-24
Step 7: P-1 Model Definition, Solution, and Postprocessing

You will now repeat the above calculation using the P-1 radiation model. The main steps
are identical to the procedure described above for the Rosseland model.
1. Enable the P-1 model.

2. Confirm that the wall emissivity is 1 for all walls.

For each wall boundary, there will be a new entry, Internal Emissivity, in the Thermal
section of the Wall panel. Retain the default value of 1.
3. Modify the under-relaxation parameters.

(a) Under Under-Relaxation Factors, set the factor for P1 to 1.0, and retain the
default factors for Pressure, Momentum, and Energy (0.3, 0.7, and 1.0).
Note that an additional equation, P1, appears because the P-1 model solves
an additional radiation transport equation. This problem is relatively easy
to converge for the P-1 model since there is not much coupling between the
radiation and temperature equations at low optical thicknesses. Consequently
a high under-relaxation factor can be used for P-1.
4. Save the case file (rad p1.cas).


The P-1 model reaches convergence after about 115 additional iterations.
6. Save the data file (rad p1.dat).


7. Examine the results of the P-1 model calculation.

Note: The steps below do not include detailed instructions because the procedure
is the same one that you followed for the Rosseland model postprocessing.
See Step 6: Postprocessing for the Rosseland Model if you need more
detailed instructions.
(a) Display velocity vectors (Figure 5.9).
2.86e-04
2.72e-04
2.58e-04
2.43e-04
2.29e-04
2.15e-04
2.00e-04
1.86e-04
1.72e-04
1.58e-04
1.43e-04
1.29e-04
1.15e-04
1.00e-04
8.60e-05
7.17e-05
5.74e-05
4.31e-05
2.88e-05
1.45e-05
1.98e-07

Figure 5.9: Velocity Vectors for the P-1 Model
(b) Plot the y velocity along the horizontal centerline (Figure 5.10), and save the
plot data to a file called rad p1.xy.
You will need to reselect Y Velocity under Y Axis Function. Also, remember
! to turn off the Write to File option so that you can access the Plot button
to generate the plot.
(c) Compute the total wall heat transfer rate.
Report −→Fluxes ...
The total heat transfer rate reported on the right wall is 8.47 × 105 W. The
heat imbalance at the lateral walls is negligibly small. You will see later that
the Rosseland and P-1 wall heat transfer rates are substantially different from
those obtained by the DTRM and the DO model.
5-26
y=0.5
2.50e-04
2.00e-04
1.50e-04
1.00e-04
5.00e-05
0.00e+00
Y
Velocity -5.00e-05
(m/s) -1.00e-04
-1.50e-04
-2.00e-04
-2.50e-04
-3.00e-04
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Position (m)
Y Velocity
Figure 5.10: XY Plot of Centerline y Velocity for the P-1 Model
Notice how different the velocity vectors and y-velocity profile are from those obtained
using the Rosseland model. The P-1 velocity profiles show a clear momentum boundary
layer along the hot and cold walls. These profiles are much closer to those obtained from
the non-radiating case (Figures 5.11 and 5.12). Though the P-1 model is not appropriate
for this optically thin limit, it yields the correct velocity profiles since the radiation source
in the energy equation, which is proportional to the absorption coefficient, is small. The
Rosseland model uses an effective conductivity to account for radiation, and yields the
wrong temperature field, which in turn results in an erroneous velocity field.

2.16e-04
2.05e-04
1.94e-04
1.83e-04
1.72e-04
1.62e-04
1.51e-04
1.40e-04
1.29e-04
1.19e-04
1.08e-04
9.70e-05
8.62e-05
7.55e-05
6.47e-05
5.39e-05
4.31e-05
3.24e-05
2.16e-05
1.08e-05
3.92e-08

Figure 5.11: Velocity Vectors with No Radiation
y=0.5
2.50e-04
2.00e-04
1.50e-04
1.00e-04
5.00e-05
Y 0.00e+00
Velocity
-5.00e-05
(m/s)
-1.00e-04
-1.50e-04
-2.00e-04
-2.50e-04
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Position (m)
Y Velocity
Figure 5.12: XY Plot of Centerline y Velocity with No Radiation
5-28
Step 8: DTRM Definition, Solution, and Postprocessing

1. Turn on the discrete transfer radiation model (DTRM) and define the ray tracing.
(a) Select Discrete Transfer under Model.

(b) Accept the defaults by clicking OK.
The Ray Tracing panel will open automatically.
(c) Accept the default settings for Clustering and Angular Discretization by clicking
OK.
When you click OK, FLUENT will open a Select File dialog box so you can
specify a name for the ray file used by the DTRM. The number of Cells Per
Volume Cluster and Faces Per Surface Cluster control the total number of radiat-
ing surfaces and absorbing cells. For a small 2D problem, the default number
of 1 is acceptable. For a large problem, however, you will want to increase
these numbers to reduce the ray tracing expense. The Theta Divisions and Phi
Divisions control the number of rays being created from each surface cluster.
For most practical problems, the default settings will suffice.

See Section 12.3.3 of the User’s Guide for a more detailed description of the
ray tracing procedure.
(d) In the Ray File text entry box in the Select File dialog box, enter rad dtrm.ray
for the name of the ray file. Then click OK.
FLUENT will print an informational message describing the progress of the ray
tracing procedure.
2. Retain the current under-relaxation factors for pressure, momentum, and energy
(0.3, 0.7, and 1.0).
3. Save the case file (rad dtrm.cas).


The solution will converge after about 80 additional iterations.
5. Save the data file (rad dtrm.dat).

6. Examine the results of the DTRM calculation.

plot data to a file called rad dtrm.xy.
5-30
2.88e-04
2.73e-04
2.59e-04
2.44e-04
2.30e-04
2.16e-04
2.01e-04
1.87e-04
1.73e-04
1.58e-04
1.44e-04
1.30e-04
1.15e-04
1.01e-04
8.64e-05
7.21e-05
5.77e-05
4.33e-05
2.90e-05
1.46e-05
2.42e-07

Figure 5.13: Velocity Vectors for the DTRM
y=0.5
2.50e-04
2.00e-04
1.50e-04
1.00e-04
5.00e-05
0.00e+00
Y
Velocity -5.00e-05
(m/s) -1.00e-04
-1.50e-04
-2.00e-04
-2.50e-04
-3.00e-04
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Position (m)
Y Velocity
Figure 5.14: XY Plot of Centerline y Velocity for the DTRM


The total heat transfer rate reported on the right wall is 6.06 × 105 W. Note
that this is substantially lower than the values predicted by the Rosseland and
P-1 models.
5-32
Step 9: DO Model Definition, Solution, and Postprocessing

1. Turn on the discrete ordinates (DO) radiation model and define the angular dis-
cretization.
(a) Select Discrete Ordinates under Model.

The panel will expand to show additional inputs for the DO model.
(b) Set the number of Flow Iterations Per Radiation Iteration to 1.
This is a relatively simple flow problem, and will converge easily. Consequently
it is useful to do the DO calculation every iteration of the flow solution. For
problems that are difficult to converge, it is sometimes useful to allow the flow
solution to establish itself between radiation calculations. In such cases, it may
be useful to set Flow Iterations Per Radiation Iteration to a higher value, such
as 10.
(c) Retain the default settings for Angular Discretization and Non-Gray Model.
The Number of Bands for the Non-Gray Model is zero because only gray radia-
tion is being modeled in this tutorial.
See Section 12.3.6 of the User’s Guide for details about the angular discretiza-
tion used by the DO model.
Note: When you click OK in the Radiation Model panel, FLUENT will present
an Information dialog box telling you that new material properties have
been added for the radiation model. The property that is new for the DO
model is the refractive index, which is relevant only when you are model-
ing semi-transparent media. Since you are not modeling semi-transparent
media here, you can simply click OK in the dialog box to acknowledge this
information.

(0.3, 0.7, and 1.0), as well as the default under-relaxation of 1 for the discrete
ordinates transport equation.
3. Save the case file (rad do.cas).


5. Save the data file (rad do.dat).

6. Examine the results of the DO calculation.

2.89e-04
2.75e-04
2.60e-04
2.46e-04
2.31e-04
2.17e-04
2.03e-04
1.88e-04
1.74e-04
1.59e-04
1.45e-04
1.30e-04
1.16e-04
1.01e-04
8.69e-05
7.25e-05
5.80e-05
4.36e-05
2.91e-05
1.47e-05
2.23e-07

Figure 5.15: Velocity Vectors for the DO Model
5-34
plot data to a file called rad do.xy.
y=0.5
3.00e-04
2.00e-04
1.00e-04
Y 0.00e+00
Velocity
(m/s)
-1.00e-04
-2.00e-04
-3.00e-04
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Position (m)
Y Velocity
Figure 5.16: XY Plot of Centerline y Velocity for the DO Model

The total heat transfer rate reported on the right wall is 6.12 × 105 W. Note
that this is about 1.5% higher than that predicted by the DTRM. The DO
and DTRM values are comparable to each other, while the Rosseland and P-1
values are both substantially different. The DTRM and DO models are valid
across the range of optical thickness, and the heat transfer rates computed
using them are expected to be closer to the correct heat transfer rate.

Step 10: Comparison of y-Velocity Plots

In this step, you will read the plot files you saved for all the solutions and compare them
in a single plot.
Plot −→File...
1. Read in all the XY plot files.

(a) Click on the Add... button.
(b) In the resulting Select File dialog box, select rad do.xy, rad dtrm.xy, rad p1.xy,
and rad ross.xy in the Files list.
They will be added to the XY File(s) list. If you accidentally add an incorrect
file, you can select it in this list and click Remove.
(c) Click OK to load the 4 files.
2. Click on Plot.
Extra: You can click Curves... in the File XY Plot panel to open the Curves panel,
where you can define different styles for different plot curves. In Figure 5.17,
different symbols have been selected for each curve.
3. Resize and move the legend box so that you can read the information inside it.
(a) To resize the box, press any mouse button on a corner and drag the mouse to
the desired position.
(b) To move the legend box, press any mouse button anywhere else on the box
and drag it to the desired location.
5-36
Y Velocity
Y Velocity
3.00e-04
Y Velocity (rad_dtrm.xy)
Y Velocity (rad_p1.xy)
Y Velocity (rad_ross.xy)
2.00e-04
1.00e-04
Y 0.00e+00
Velocity
-1.00e-04
-2.00e-04
-3.00e-04
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Position
Y Velocity
Figure 5.17: Comparison of Computed y Velocities for aL = 0.2
Notice in Figure 5.17 that the velocity profiles for the P-1 model, DTRM, and DO model
are nearly identical even though the reported wall heat transfer rates are different. This
is because in an optically thin problem, the velocity field is essentially independent of the
radiation field, and all three models give a flow solution very close to the non-radiating
case. The Rosseland model gives substantially erroneous solutions for an optically thin
case.

Step 11: Comparison of Radiation Models for an Optically Thick Medium

In the previous steps, you compared the results of four radiation models for an optically
thin (aL = 0.2) medium. It was found that, as a result of the low optical thickness,
the velocity fields predicted by the P-1, DTRM, and DO models were very similar and
close to that obtained in the non-radiating case. The wall heat transfer rates for DO
and DTRM were very close to each other, and substantially different from those obtained
with the Rosseland and P-1 models. In this step, you will recalculate a solution (using
each radiation model) for an optically thick (aL = 5) medium. This is accomplished by
increasing the value of the absorption coefficient from 0.2 to 5. You will repeat the process
outlined below for each set of case and data files that you saved earlier in the tutorial.
1. For each radiation model, calculate a new solution for aL = 5.

(a) Read in the case and data file saved earlier (e.g.rad ross.cas and
rad ross.dat).
(b) Set the absorption coefficient to 5.
This will result in an optical thickness aL of 5, since L = 1.
(c) Calculate until the new solution converges.
For the DTRM calculation, you may need to click the Iterate button re-
! peatedly until the radiation field is updated. Since the number of Flow
Iterations Per Radiation Iteration in the Radiation Model panel is 10, it is
possible that the radiation field will not be updated for as many as 9 it-
erations, although FLUENT will report that the solution is converged. If
this happens, keep clicking the Iterate button until the radiation field is
updated and the solution proceeds for multiple iterations.
(d) Save the new case and data files using a different file name (e.g., rad ros5.cas
and
rad ros5.dat).
(e) Compute the total wall heat transfer rate.
(f) Plot the y velocity along the horizontal centerline, and save the plot data to
a file (e.g., rad ros5.xy).
5-38
2. Compare the computed heat transfer rates for the four models.
The wall heat transfer rates predicted by the four radiation models range from 3.50×
105 to 3.97 × 105 W.
3. Compare the y-velocity profiles in a single plot (Figure 5.18).

Plot −→File...
Note: Use the Delete button in the File XY Plot panel to remove the old XY plot
data files.
Y Velocity
Y Velocity
Y Velocity (rad_do5.xy)
5.00e-04
Y Velocity (rad_p15.xy)
Y Velocity (rad_ross5.xy)
4.00e-04
3.00e-04
2.00e-04
1.00e-04
Y 1.36e-20
Velocity
-1.00e-04
-2.00e-04
-3.00e-04
-4.00e-04
-5.00e-04
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Position
Y Velocity
Figure 5.18: Comparison of Computed y Velocities for aL = 5
The XY plots of y velocity are nearly identical for the P-1 model, DO model, and
DTRM. The Rosseland model gives somewhat different velocities, but is still within
10% of the other results. The Rosseland and P-1 models are suitable for the opti-
cally thick limit; the DTRM and DO models are valid across the range of optical
thicknesses. Consequently, they yield similar answers at aL = 5. For many ap-
plications with large optical thicknesses, the Rosseland and P-1 models provide a
simple low-cost alternative.

Step 12: S2S Model Definition, Solution and Postprocessing for a

Non-Participating Medium
In the previous steps, you compared the results of four radiation models for optically thin
(aL = 0.2) and optically thick (aL = 5) media.
The surface-to-surface (S2S) radiation model can be used to account for the radiation ex-
change in an enclosure of gray-diffuse surfaces. The energy exchange between two surfaces
depends in part on their size, separation distance, and orientation. These parameters are
accounted for by a geometric function called a “view factor”.
The S2S model assumes that all surfaces are gray and diffuse. Thus, according to the
gray-body model, if a certain amount of radiation is incident on a surface, a fraction is
reflected, a fraction is absorbed, and a fraction is transmitted. The main assumption of
the S2S model is that any absorption, emission, or scattering of radiation by the medium
can be ignored; therefore, only ”surface-to-surface” radiation needs to be considered for
analysis.
For most applications the surfaces in question are opaque to thermal radiation (in the
infrared spectrum), so the surfaces can be considered opaque. For gray, diffuse, and
opaque surfaces it is valid to assume that the emissivity is equal to the absorptivity and
that reflectivity is equal to 1 minus the emissivity.
When the S2S model is used, you also have the option to define a “partial enclosure”; i.e.,
you can disable the view factor calculation for walls with negligible emission/absorption
or walls that have uniform temperature. The main advantage is to speed up the view
factor calculation and the radiosity calculation. In this step, you will calculate a solution
for aL = 0 using the S2S radiation model without partial enclosure. In the next step,
you will use the DTRM and DO models for aL = 0, and compare the results of the three
models. The Rosseland and P-1 models are not considered here as they have been shown
(earlier in the tutorial) to be inappropriate for optically thin media. Later in the tutorial,
you will calculate a solution for S2S model with partial enclosure and compare the results
with the solution for S2S model for a non-participating medium calculated here.
1. Turn on the surface-to-surface (S2S) radiation model and define the view factor
and cluster parameters.
(a) Select Surface to Surface under Model.
The panel will expand to show additional inputs for the S2S model.
5-40
(b) Set the view factor and cluster parameters.

i. Click Set... under Parameters.
The View Factor and Cluster Parameters panel will open automatically.
ii. Click OK to accept the default settings.

The S2S radiation model is computationally very expensive when there are
a large number of radiating surfaces. The number of radiating surfaces is

reduced by clustering surfaces into surface “clusters”. The surface clusters

are made by starting from a face and adding its neighbors and their neigh-
bors until a specified number of faces per surface cluster is collected. For
a small 2D problem, the default value of 1 for Faces Per Surface Cluster is
acceptable. For a large problem, you can increase this number to reduce
the memory requirement for the view factor file that is saved in a later
step. This may also lead to some reduction in the computational expense.
However, this is at the cost of some accuracy.
Using the Blocking option ensures that any additional surface that is block-
ing the view between two opposite surfaces is considered in the view factor
calculation. In this case, there is no obstructing surface between the op-
posite walls, so selecting either the Blocking or the Nonblocking option
will produce the same result. The default setting for Smoothing is None,
which is appropriate for small problems. The Least Square option is more
accurate, but also more computationally expensive.
See Section 12.3.13 of the User’s Guide for details about view factors and
clusters for the S2S model.
(c) Compute the view factors for the S2S model.
This step is required only if the problem is being solved for the first time. For
subsequent calculations, you can read the view factor and cluster information
from an existing file (by clicking Read... instead of Compute/Write...).
i. Click Compute/Write... under Methods in the Radiation Model panel.
FLUENT will open a Select File dialog box so you can specify a name for
the file where the cluster and view factor parameters are stored.
ii. In the S2S File text entry box in the Select File dialog box, enter rad s2s.s2s
for the name of the S2S file. Then click OK.
FLUENT will print an informational message describing the progress of
the view factor calculation.
The size of the viewfactor file can be very large if not compressed. It is
i highly recommended to compress the view factor file by providing .gz or
.Z extension after the name (i.e. rad s2s.gz or rad s2s.Z).
(0.3, 0.7, and 1.0).
3. Save the case file (rad s2s.cas).

5-42

5. Save the data file (rad s2s.dat).

6. Examine the results of the S2S calculation.

2.48e-04
2.36e-04
2.23e-04
2.11e-04
1.99e-04
1.86e-04
1.74e-04
1.61e-04
1.49e-04
1.37e-04
1.24e-04
1.12e-04
9.94e-05
8.70e-05
7.46e-05
6.22e-05
4.98e-05
3.74e-05
2.50e-05
1.26e-05
2.29e-07

Figure 5.19: Velocity Vectors for the S2S Model
plot data to a file called rad s2s.xy.
You will have to reselect Y Velocity under Y Axis Function. Also, remember
! to turn off the Write to File option to access the Plot button to generate
the plot.

y=0.5
2.50e-04
2.00e-04
1.50e-04
1.00e-04
5.00e-05
Y 0.00e+00
Velocity
-5.00e-05
(m/s)
-1.00e-04
-1.50e-04
-2.00e-04
-2.50e-04
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Position (m)
Y Velocity
Figure 5.20: XY Plot of Centerline y Velocity for the S2S Model

The total heat transfer rate on the right wall is 6.77 × 105 W.
5-44
Step 13: Comparison of Radiation Models for a Non-Participating Medium

In this step, you will calculate a solution for the aL = 0 case, using the DTRM and DO
models, and then compare the results with the S2S results.
1. For the DTRM and DO models, calculate a new solution for aL = 0.

(a) Read in the case and data files saved earlier (e.g., rad dtrm.cas and
rad dtrm.dat).
(b) Set the absorption coefficient to 0.
This will result in an optical thickness aL of 0.
(c) Calculate until the new solution converges.
For the DTRM calculation, you may need to click the Iterate button re-
! peatedly until the radiation field is updated. Since the number of Flow
Iterations Per Radiation Iteration in the Radiation Model panel is 10, it is
possible that the radiation field will not be updated for as many as 9 it-
erations, although FLUENT will report that the solution is converged. If
this happens, keep clicking the Iterate button until the radiation field is
updated and the solution proceeds for multiple iterations.
(d) Save the new case and data files using a different file name (e.g., rad dtr0.cas
and rad dtr0.dat).
(e) Compute the total wall heat transfer rate.
(f) Plot the y velocity along the horizontal centerline, and save the plot data to
a file (e.g., rad dtr0.xy)
2. Compare the computed heat transfer rates for the three models.
For the S2S model, the total heat transfer rate on the right wall was 6.77 × 105 W.
This is about 5% higher than that predicted by the DTRM and 1.5% higher than DO.
Although the S2S, DO, and DTRM values are comparable to each other, this problem
involves enclosure radiative transfer without participating media. Therefore, the
S2S model provides the most accurate solution.

3. Compare the y-velocity profiles in a single plot (Figure 5.21)

Plot −→File...
(a) Use the Delete button in the File XY Plot panel to remove the old XY plot
data files.
(b) Read in all the XY plot files you saved for the S2S, DTRM, and DO models.
(c) Click on Plot.
Y Velocity
Y Velocity
3.00e-04
Y Velocity (rad_dtrm.xy)
Y Velocity (rad_dtr0.xy)
2.00e-04
1.00e-04
Y 0.00e+00
Velocity
-1.00e-04
-2.00e-04
-3.00e-04
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Position
Y Velocity
Figure 5.21: Comparison of Computed y Velocities for aL = 0
In Figure 5.21, the velocity profiles for the DTRM, DO, and S2S models are almost
identical even though the wall heat transfer rates are different.
5-46
Step 14: S2S Model Definition, Solution and Postprocessing with Partial
Enclosure
As mentioned earlier, when the S2S model is used, you also have the option to define
a ”partial enclosure”; i.e., you can disable the view factor calculation for walls with
negligible emission/absorption, or walls that have uniform temperature. Even though the
view factor will not be computed for these walls, they will still emit radiation at a fixed
temperature called the ”partial enclosure temperature”. The main advantage of this is to
speed up the view factor and the radiosity calculation.
For this problem, specify the left wall boundary as the non-participating wall in S2S
radiation. Consequently, you need to specify the partial enclosure temperature for the wall
boundary that is not participating in S2S radiation. Note that if multiple wall boundaries
are not participating in S2S radiation and each has a different temperature, then the
partial enclosure option may not yield accurate results. This is because the same partial
enclosure temperature is specified for each of the non-participating walls.
1. Read in the case and data files for the S2S model (rad s2s.cas and rad s2s.dat).
2. In the Radiation Model panel, retain Surface to Surface (S2S) as the radiation model.
3. Under Partial Enclosure, set the Temperature to 1000 k.

Previous radiation model setups for this problem specified left wall temperature as
1000 k. Therefore, set the partial enclosure temperature to this temperature.
4. Click OK to close the Radiation Model panel.

5. Set the boundary conditions for the left wall.

Define −→Boundary Conditions
(a) Turn off the Participates in S2S Radiation option.

You will now revisit the Radiation Model panel to recompute the view factors.
6. Compute the view factors for the S2S model.

The view factor file will store the view factors for the radiating surfaces only. This
may help you to control the size of the view factor file as well as the memory
required to store view factors in FLUENT. Furthermore, the time required to compute
the view factors will reduce as only the view factors for radiating surfaces will be
calculated.
You should compute the view factors only when you have specified the
! boundaries that will participate in the radiation model using the Boundary
Conditions panel. If you first compute the view factors and then make a
change to the boundary conditions, FLUENT will use the view factor file
stored earlier for calculating a solution, in which case, the changes that
you made to the model will not be used for the calculation. Therefore,
you should recompute the view factors and save the case file whenever you
modify the number of objects that will participate in radiation.
5-48
(a) Click Compute/Write... under Methods in the Radiation Model panel.

FLUENT will open a Select File dialog box so you can specify a name for the
file where the cluster and view factor parameters are stored.
(b) In the S2S File text entry box in the Select File dialog box, enter rad s2sp.s2s
for the name of the S2S file. Then click OK.
FLUENT will print an informational message describing the progress of the
view factor calculation.
(0.3, 0.7, and 1.0).
8. Save the case file (rad s2sp.cas).


10. Save the data file (rad s2sp.dat).

11. Examine the results of the S2S calculation.

plot data to a file called rad s2sp.xy.
You will have to reselect Y Velocity under Y Axis Function. Also, remember
! to turn off the Write to File option to access the Plot button to generate
the plot.

2.47e-04
2.35e-04
2.23e-04
2.10e-04
1.98e-04
1.85e-04
1.73e-04
1.61e-04
1.48e-04
1.36e-04
1.24e-04
1.11e-04
9.90e-05
8.67e-05
7.43e-05
6.20e-05
4.96e-05
3.73e-05
2.49e-05
1.26e-05
2.27e-07

Figure 5.22: Velocity Vectors for the S2S Model with Partial Enclosure
y=0.5
2.50e-04
2.00e-04
1.50e-04
1.00e-04
5.00e-05
Y 0.00e+00
Velocity
-5.00e-05
(m/s)
-1.00e-04
-1.50e-04
-2.00e-04
-2.50e-04
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Position (m)
Y Velocity
Figure 5.23: XY Plot of Centerline y Velocity for the S2S Model with Partial Enclosure
5-50
The total heat transfer rate on the right wall is 6.77 × 105 W. Note that the
total heat transfer rate on the left wall is reported as zero because the report
utility in the current version of FLUENT does not account for the radiation
heat transfer rate by this wall, as it should.

Step 15: Comparison of S2S Models with and without Partial Enclosure
1. Compare the computed heat transfer rates for the two S2S models.
2. Compare the y-velocity profiles in a single plot (Figure 5.24)

Plot −→File...
(a) Use the Delete button in the File XY Plot panel to remove the old XY plot
data files.
(b) Read in all the XY plot files you saved for the S2S models.
(c) Click on Plot.
Y Velocity
Y Velocity
Y Velocity (rad_s2sp.xy)
2.50e-04
2.00e-04
1.50e-04
1.00e-04
5.00e-05
Y 0.00e+00
Velocity
-5.00e-05
-1.00e-04
-1.50e-04
-2.00e-04
-2.50e-04
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Position
Y Velocity
Figure 5.24: Comparison of Computed y Velocities for S2S models
In Figure 5.24, the velocity profiles for the S2S model without partial enclosure and the
S2S model with partial enclosure are almost identical.
5-52
Summary
In this tutorial, you studied combined natural convection and radiation in a square box
and compared the performance of four radiation models in FLUENT for optically thin
and optically thick cases, and the performance of three radiation models for a non-
participating medium.
• For the optically thin case, the Rosseland and P-1 models are not appropriate; the
DTRM and the DO model are applicable, and yield similar results.
• In the optically thick limit, all four models are appropriate and yield similar results.
In this limit, the less computationally-expensive Rosseland and P-1 models may be
adequate for many engineering applications.
• The S2S radiation model is appropriate for modeling the enclosure radiative transfer
without participating media, where the methods for participating radiation may not
always be efficient.
See Section 12.3 of the User’s Guide for more information about the applicability of the
different radiation models.

5-54
Tutorial 6. Using a Non-Conformal Mesh
Introduction
Film cooling is a process that is used to protect turbine vanes in a gas turbine engine
from exposure to hot combustion gases. This tutorial illustrates how to set up and solve
a film cooling problem using a non-conformal mesh. The system that is modeled consists
of three parts: a duct, a hole array, and a plenum. The duct is modeled with a hexahedral
mesh, and the plenum and hole regions are modeled using a tetrahedral mesh. These two
meshes are merged together to form a “hybrid” mesh, with a non-conformal interface
boundary between them.
Due to symmetry of the hole array, only a portion of the geometry is modeled in FLUENT,
with symmetry applied to the outer boundaries. The duct contains a high-velocity fluid in
streamwise flow (Figure 6.1). An array of holes intersects the duct at an inclined angle,
and a cooler fluid is injected into the holes from a plenum. The coolant that moves
through the holes acts to cool the surface of the duct, downstream of the injection. Both
fluids are air, and the flow is classified as turbulent. The velocity and temperature of the
streamwise and cross-flow fluids are known, and FLUENT is used to predict the flow and
temperature fields that result from convective heat transfer.
• Merge hexahedral and tetrahedral meshes to form a hybrid mesh
• Create a non-conformal grid interface
• Model heat transfer across a non-conformal interface with specified temperature

and velocity boundary conditions
• Plot temperature profiles on specified isosurfaces
Prerequisites

Using a Non-Conformal Mesh
Problem Description
This problem considers a model of a 3D section of a film cooling test rig. A schematic is
shown in Figures 6.1 and 6.2. The problem consists of a duct, 24.5 in long, with cross-
sectional dimensions of 0.75 in × 5 in. An array of uniformly spaced holes is located at
the bottom of the duct. Each hole has a diameter of 0.5 inches, is inclined at 35 degrees,
and is spaced 1.5 inches apart laterally. Cooler injected air enters the system through
the plenum, with cross-sectional dimensions of 3.3 in × 1.25 in.
Because of the symmetry of the geometry, only a portion of the domain needs to be
modeled. The computational domain is shown in outline in Figure 6.2. The bulk tem-
perature of the streamwise air (T∞ ) is 273 K, and the velocity of the air stream is 20
m/s. The bottom wall of the duct that intersects the hole array is assumed to be a
completely insulated (adiabatic) wall. The secondary (injected) air enters the plenum at
a uniform velocity of 0.4559 m/s. The temperature of the injected air (Tinject ) is 136.6 K.
The properties of air that are used in the model are shown in Figure 6.2.
9.5 in 0.5 in 14.5 in
v = 20 m/s
Τ
∞ = 273 K Duct
y 5 in
x
1.25 in Hole 35°
1.25 in Plenum
3.3 in
v = 0.4559 m/s
T = 136.6 K
inject
Figure 6.1: Schematic of the Problem, Front View
6-2
z
Τ
∞ = 273 K
0.25 in 0.75 in
µ = 0.000017894 kg/m-s
c p = 1006.43 J/kg-K
Figure 6.2: Schematic of the Problem, Top View
Setup and Solution

Preparation
1. Download non_conformal_mesh.zip from the Fluent Inc. User Services Center
2. Unzip non_conformal_mesh.zip.
film hex.msh and film tet.msh can be found in the /non conformal mesh folder
created after unzipping the file.

Step 1: Merging the Mesh Files

1. Start the 3D version of tmerge by typing utility tmerge -3d at the system
prompt.
2. Provide the mesh file names film tet.msh and film hex.msh as prompted. Pro-
vide scaling of 1 and translations and rotations of zero for each mesh file. Save the
new merged mesh file as filmcool.msh.
Append 3D grid files.

tmerge3D Fluent Inc, Version 2.1.11
Enter name of grid file (ENTER to continue) : film_tet.msh
x,y,z scaling factor, eg. 1 1 1 : 1 1 1
x,y,z translation, eg. 0 1 0 : 0 0 0
rot axis (0,1,2), angle (deg), eg. 1 45 : 0 0
Enter name of grid file (ENTER to continue) : film_hex.msh
x,y,z scaling factor, eg. 1 1 1 : 1 1 1
x,y,z translation, eg. 0 1 0 : 0 0 0
rot axis (0,1,2), angle (deg), eg. 1 45 : 0 0
Enter name of grid file (ENTER to continue) : <ENTER>
Enter name of output file : filmcool.msh
The mesh files must be read into tmerge in this order for the tutorial to run
! as written. Otherwise, zone names and numbers will be assigned differently
when the files are merged together. In general, however, you can specify
files to be read into tmerge in any order.
6-4
Step 2: Grid
2. Read in the mesh file filmcool.msh.

3. Check the grid.

Grid −→Check
4. Scale the grid and change the unit of length to inches.

Grid −→Scale...
(a) In the Unit Conversion drop-down list, select in to complete the phrase Grid
Was Created In in (inches).
(c) Click Change Length Units to set inches as the working units for length.
The final Domain Extents should appear as in the panel above.
(d) Close the panel.

5. Display an outline of the 3D grid (Figure 6.3).

(a) In the Surfaces list, deselect symmetry-3, symmetry-5 and symmetry-tet.

(b) Click Display.
Y
Z X
Grid
Figure 6.3: Hybrid Mesh for Film Cooling Problem
6-6
(c) Zoom in using your middle mouse button to get the view displayed in Fig-
ure 6.4.
Y
Z X
Grid
Figure 6.4: Hybrid Mesh (Zoomed-In View)
In Figure 6.4 you can see the quadrilateral faces of the hexahedral cells that are
used to model the duct region, and the triangular faces of the tetrahedral cells that
are used to model the plenum and hole regions, resulting in a hybrid mesh.
Extra: You can use the right mouse button to check which zone number corre-
sponds to each boundary. If you click the right mouse button on one of the
boundaries in the graphics window, its zone number, name, and type will be
printed in the FLUENT console window. This feature is especially useful when
you have several zones of the same type and you want to distinguish between
them quickly.
6. Close the Grid Display panel.

Step 3: Models

6-8
3. Enable the standard k- turbulence model.

(a) Activate k-epsilon (2 eqn) under Model to expand the panel.

(b) Click OK.

Step 4: Materials
1. Retain air as the Fluent Fluid Materials , and use the incompressible-ideal-gas law to
compute density. Retain the default values for all other properties.
Do not forget to click the Change/Create button after selecting

!
incompressible-ideal-gas in the drop-down list for Density.
6-10

1. Keep the default operating conditions.


1. Set the boundary conditions for the streamwise flow inlet (velocity-inlet-1).
(a) Change the Zone Name from velocity-inlet-1 to velocity-inlet-duct.

(b) Set the Velocity Magnitude to 20 m/s.
(d) In the Turbulence Specification Method drop-down list, select Intensity and Hy-
draulic Diameter.
(e) Set the Turbulence Intensity to 1% and the Hydraulic Diameter to 5 in.
2. Set the boundary conditions for the injected stream inlet (velocity-inlet-14).
(a) Change the Zone Name from velocity-inlet-14 to velocity-inlet-plenum.
(b) Set the Velocity Magnitude to 0.4559 m/s.
(c) Set the Temperature to 136.6 K.
(d) In the Turbulence Specification Method drop-down list, select Intensity and Vis-
cosity Ratio.
6-12
(e) Set the Turbulence Intensity to 1% and keep the Turbulent Viscosity Ratio default
of 10.
In the absence of any identifiable length scale for turbulence, the Intensity and Vis-
cosity Ratio method should be used.
See Chapter 11 of the User’s Guide for more information on how to set the boundary
conditions for turbulence.
3. Set the boundary conditions for the flow exit (pressure-outlet-1).

(a) Change the Zone Name from pressure-outlet-1 to pressure-outlet-duct.

(b) Keep the default setting of 0 Pa for Gauge Pressure.
(c) Set the Backflow Total Temperature to 273 K.
(d) In the Turbulence Specification Method drop-down list, select Intensity and Vis-
cosity Ratio.
(e) Set the Backflow Turbulence Intensity to 1% and keep the Backflow Turbulent
Viscosity Ratio default of 10.
4. Set the conditions for the fluid in the duct (fluid-9).
6-14
(a) Change the Zone Name from fluid-9 to fluid-duct.

(b) Keep the default selection of air as the Material Name.

5. Set the conditions for the fluid in the plenum and hole (fluid-17).
(a) Change the Zone Name from fluid-17 to fluid-plenum.
(b) Keep the default selection of air as the Material Name.
6-16
6. Keep the default boundary conditions for the plenum and hole wall (wall-15).

7. Define the zones on the non-conformal boundary as interface zones.

The non-conformal grid interface contains two boundary zones: wall-1 and wall-12.
wall-1 is the bottom surface of the duct, and wall-12 represents the hole through
which the cool air is injected from the plenum (Figure 6.5). These boundaries were
defined as walls in the original mesh files, film hex.msh and film tet.msh, and
must be redefined as interface boundary types.
(a) Open the Grid Display panel.
(b) Select wall-1 and wall-12 under Surfaces.
(c) Click Display.
i. Activate bottom view.
A. Select bottom under Views and click Restore.
Zoom in using middle mouse button to get the view as shown below.
X Y
Z
Grid
Figure 6.5: Grid for the wall-1 and wall-12 Boundaries
6-18
8. Change the Type for wall-1 and wall-12 to interface.

(a) Select wall-1 in the Zone list and choose interface as the new Type.
(b) Click Yes when asked OK to change wall-1’s type from wall to interface?.
(c) Change the Zone Name to interface-duct.
(d) Repeat this procedure to convert wall-12 to an interface boundary zone named
interface-hole.
9. Close the Boundary Conditions panel.

Step 7: Grid Interfaces

In this step, you will create a non-conformal grid interface between the hexahedral and
tetrahedral meshes.
Define −→Grid Interfaces...
1. Select interface-hole in the Interface Zone 1 list.
When one interface zone is smaller than the other, it is recommended that
! you choose the smaller zone as Interface Zone 1.
2. Select interface-duct in the Interface Zone 2 list.
3. Enter the name junction under Grid Interface.
4. Click Create.
Note: In the process of creating the grid interface, FLUENT creates two new wall
boundary zones: wall-11 and wall-18. You will not be able to display these
walls.
wall-11 is the non-overlapping region of the interface-hole zone that results from
the intersection of the interface-hole and interface-duct boundary zones, and is
listed under Boundary Zone 1 in the Grid Interfaces panel. wall-11 is empty,
since interface-hole is completely contained within the interface-duct boundary.
wall-18 is the non-overlapping region of the interface-duct zone that results from
the intersection of the two interface zones, and is listed under Boundary Zone
2 in the Grid Interfaces panel.
6-20
In general, you will need to set boundary conditions for wall-18 (since it is
!
not empty). In this case, default settings are used.
Step 8: Solution
(a) Under Discretization, select Second Order Upwind for Momentum and Turbulence
Kinetic Energy.
(b) Scroll down the list and select Second Order Upwind for Turbulence Dissipation
Rate and Energy.

(b) Click the OK button.


(a) Select velocity-inlet-duct in the Compute From drop-down list.

6-22
4. Save the case file (filmcool.cas).


(a) Set the Number of Iterations to 250.

(b) Click Iterate.
Note: During the first few iterations, the console window will report that turbu-
lent viscosity is limited in a couple of cells. This message should go away as
the solution converges and the turbulent viscosity approaches more reasonable
levels.
The solution will converge after about 135 iterations.
6. Save the case and data files (filmcool.cas and filmcool.dat).

Note: If you choose a file name that already exists in the current directory, FLU-
ENT will prompt you for confirmation to overwrite the file.

1. Display filled contours of static pressure (Figure 6.6).
(a) Select Filled under Options.

(b) Select Pressure... and Static Pressure in the Contours of drop-down lists.
(c) In the Surfaces list, select interface-duct and interface-hole.
(d) Scroll down the Surfaces list and select symmetry-1, symmetry-tet, and wall-15.
6-24
(e) Reset the view to the default view.

i. Click Default under Actions.
ii. Close the panel.

(f) In the Contours panel, click Display.
(g) Zoom in on the view to get the display shown in Figure 6.6.
3.36e+02
3.12e+02
2.88e+02
2.64e+02
2.39e+02
2.15e+02
1.91e+02
1.67e+02
1.42e+02
1.18e+02
9.37e+01
6.95e+01
4.52e+01
2.09e+01
-3.33e+00
-2.76e+01
-5.19e+01
-7.61e+01
-1.00e+02 Y
-1.25e+02 Z X
-1.49e+02

Note the high/low pressure zones on the upstream/downstream sides of the

coolant hole, where the jet first penetrates the primary flow in the duct.

2. Display filled contours of static temperature (Figures 6.7 and 6.8).


lists.
(b) Under Options, deselect Auto Range so that you can change the maximum and
minimum temperature gradient values to be plotted.
(c) Keep the default Min value of 0.
(d) Enter a new Max value of 273.096.
(e) Under Options, deselect Clip to Range.
(f) Click Display.
(g) Zoom in on the view to get the display shown in Figure 6.8.
Figures 6.7 and 6.8 clearly show how the coolant flow insulates the bottom of
the duct from the higher-temperature primary flow.
6-26
2.73e+02
2.59e+02
2.46e+02
2.32e+02
2.18e+02
2.05e+02
1.91e+02
1.78e+02
1.64e+02
1.50e+02
1.37e+02
1.23e+02
1.09e+02
9.56e+01
8.19e+01
6.83e+01
5.46e+01
4.10e+01
2.73e+01 Y
1.37e+01 Z X
0.00e+00

Figure 6.7: Contours of Static Temperature
2.73e+02
2.59e+02
2.46e+02
2.32e+02
2.18e+02
2.05e+02
1.91e+02
1.78e+02
1.64e+02
1.50e+02
1.37e+02
1.23e+02
1.09e+02
9.56e+01
8.19e+01
6.83e+01
5.46e+01
4.10e+01
2.73e+01 Y
1.37e+01 Z X
0.00e+00

Figure 6.8: Contours of Static Temperature (Zoomed-In View)

3. Display the velocity vectors (Figure 6.9).

(a) Select Velocity... and Velocity Magnitude in the Color by drop-down lists.
(b) Change the Scale to 2.
This will enlarge the vectors that are displayed, making it easier to view the
flow patterns.
(c) In the Surfaces list, select interface-duct and interface-hole.
(d) Scroll down the Surfaces list and select symmetry-1, symmetry-tet, and wall-15.
(e) In the Vectors panel, click Display.
(f) Zoom in on the view to get the display shown in Figure 6.9.
The flow pattern in the vicinity of the coolant hole shows the level of penetration
of the coolant jet into the main flow. Notice that the velocity field varies smoothly
across the non-conformal interface.
6-28
2.13e+01
2.02e+01
1.92e+01
1.81e+01
1.71e+01
1.60e+01
1.49e+01
1.39e+01
1.28e+01
1.17e+01
1.07e+01
9.60e+00
8.53e+00
7.47e+00
6.40e+00
5.33e+00
4.27e+00
3.20e+00
2.14e+00 Y
1.07e+00 Z X
9.16e-03

4. Plot the temperature profile along a horizontal cross-section of the duct, 0.1 inches
above the bottom.
(a) Create an isosurface on the duct surface at y = 0.1 in.
i. Select Grid... and Y-Coordinate in the Surface of Constant drop-down lists.

ii. Enter y=0.1in under New Surface Name.

iii. Enter 0.1 for Iso-Values.

iv. Click Create.
(b) Create an XY plot of static temperature on the isosurface.
i. Keep the default Plot Direction.
ii. Select Temperature... and Static Temperature in the Y-Axis Function drop-
down lists.
iii. Scroll down the Surfaces list and select y=0.1in.
iv. Click Plot.
6-30
y=0.1in
2.80e+02
2.60e+02
2.40e+02
2.20e+02
Static 2.00e+02
Temperature
(k) 1.80e+02
1.60e+02
1.40e+02
1.20e+02
-10 -7.5 -5 -2.5 0 2.5 5 7.5 10 12.5 15 17.5
Y
Z X Position (in)
Static Temperature
Figure 6.10: Static Temperature at y=0.1 in
In this plot you can see how the temperature of the fluid changes as the cool
air from the injection hole mixes with the primary flow. As expected, the tem-
perature is coolest just downstream of the hole. Note that you could also make
a similar plot on the lower wall itself, to examine the wall surface temperature.
Summary
This tutorial demonstrates how FLUENT’s non-conformal grid interface capability can
be used to handle hybrid meshes for complex geometries, such as the film cooling hole
configuration examined here. One of the principal advantages of this approach is that
it allows you to merge existing component meshes together to create a larger, more
complex mesh system, without requiring that the different components have the same
node locations on their shared boundaries. Thus, you can perform parametric studies
by merging the desired meshes, creating the non-conformal interface(s), and solving the
model. For example, in the present case, you can
• Use a different hole/plenum mesh
• Reposition the existing hole/plenum mesh
• Add additional hole/plenum meshes to create aligned or staggered multiple hole

arrays

6-32
Tutorial 7. Using a Single Rotating Reference Frame
Introduction
This tutorial considers the flow within a 2D, axisymmetric, co-rotating disk cavity system.
Understanding the behavior of such flows is important in the design of secondary air
passages for turbine disk cooling.
• Set up a 2D axisymmetric model with swirl, using a rotating reference frame
• Use the standard k- and RNG k- turbulence models with the enhanced near-wall
treatment
• Display velocity vectors and contours of pressure
• Set up and display XY plots of radial velocity
• Restart the solver from an existing solution
Prerequisites
Problem Description
The problem to be considered is shown schematically in Figure 7.1. This case is similar
to a disk cavity configuration that was extensively studied by Pincombe [1].
Air enters the cavity between two co-rotating disks. The disks are 88.6 cm in diameter
and the air enters at 1.146 m/s through a circular bore 8.86 cm in diameter. The disks,
which are 6.2 cm apart, are spinning at 71.08 rpm, and the air enters with no swirl. As
the flow is diverted radially, the rotation of the disk has a significant effect on the viscous
flow developing along the surface of the disk.

Using a Single Rotating Reference Frame
Outflow
44.3 cm Rotating Rotating

Disk Disk
6.2 cm
71.08 rpm
Inflow 4.43 cm
As noted by Pincombe [1], there are two nondimensional parameters that characterize
this type of disk cavity flow: the volume flow rate coefficient, Cw , and the rotational
Reynolds number, Reφ . These parameters are defined as follows:
Q
Cw = (7.1)
ν rout
2
Ωrout
Reφ = (7.2)
ν
where Q is the volumetric flow rate, Ω is the rotational speed, ν is the kinematic viscosity,
and rout is the outer radius of the disks. Here, you will consider a case for which Cw =
1092 and Reφ = 105 .
7-2
Setup and Solution

Preparation
1. Download single_rotating.zip from the Fluent Inc. User Services Center or copy
it from the FLUENT documentation CD to your working directory (as described in
Tutorial 1).
2. Unzip single_rotating.zip.
disk.msh can be found in the /single rotating folder created after unzipping the
file.
Step 1: Grid
1. Read the grid file (disk.msh).
2. Check the grid.

Grid −→Check


Grid
Figure 7.2: Grid Display for the Disk Cavity
7-4
Step 2: Units
1. For convenience, define new units for angular velocity and length.
In the problem description, angular velocity and length are specified in rpm and cm,
respectively. These are not the default units for these quantities.
(a) Select angular-velocity under Quantities, and rpm under Units.

(b) Select length under Quantities, and cm under Units.
(c) Close the panel.

Step 3: Models
1. Specify the solver formulation to be used for the model calculation, and enable the
modeling of axisymmetric swirl.
(a) Retain the default Segregated solver.

(b) Select Axisymmetric Swirl under Space.
(c) Retain the default selection of Absolute under Velocity Formulation.
For a rotating reference frame, the absolute velocity formulation has some
numerical advantages.
(d) Retain the default selection of Cell-Based under Gradient Option.
(e) Retain the default selection of Superficial Velocity under Porous Formulation.
7-6
2. Turn on the standard k- turbulence model with the enhanced near-wall treatment.
(a) Under Model, select k-epsilon.

The panel will expand.
(b) Keep the Standard setting under k-epsilon Model.
(c) Under Near-Wall Treatment, select Enhanced Wall Treatment and keep the de-
fault settings.
The ability to calculate a swirl velocity permits the use of a 2D mesh, so the
calculation is simpler and more economical to run. This is especially important
for problems where the enhanced wall treatment is used, and the near-wall flow
field is resolved through the viscous sublayer and buffer zones (that is, the first
grid point away from the wall is placed at a y+ on the order of 1).
See Section 11.9.3 of the User’s Guide for details.

Step 4: Materials
1. Accept the default properties for air.
For the present analysis, you will model air as an incompressible fluid with a density
of 1.225 kg/m3 and a dynamic viscosity of 1.7894×10−5 kg/m-s. Since these are
the default values, no change is required in the Materials panel.
Extra: You can modify the fluid properties for air at any time or copy another
material from the database.
See Chapter 8 of the User’s Guide for details.
7-8

You will set up the present problem using a rotating reference frame for the fluid. The
disk walls will then be defined to rotate with the moving frame.
1. Define the rotating reference frame for the fluid zone (fluid-7).
(a) Select Moving Reference Frame in the Motion Type drop-down list.
(b) Scroll down below Motion Type and set the Speed (under Rotational Velocity)
to 71.08 rpm.

2. Set the following conditions at the flow inlet (velocity-inlet-2).
3. Set the following conditions at the flow outlet (pressure-outlet-3).
Note: FLUENT will use the backflow conditions only if the fluid is flowing into
the computational domain through the outlet. Since backflow might occur at
some point during the solution procedure, you should set reasonable backflow
conditions to prevent convergence from being adversely affected.
7-10
4. Accept the default settings for the disk walls (wall-6).
Note: For a rotating reference frame, FLUENT assumes by default that all walls
rotate at the speed of the moving reference frame, and hence are moving with
respect to the stationary (absolute) reference frame. To specify a non-rotating
wall, you must specify a rotational speed of 0 in the absolute frame.

Step 6: Solution Using the Standard k- Model

(a) Under Discretization, select PRESTO! from the drop-down list to the right of
Pressure.
The PRESTO! scheme is well suited for steep pressure gradients involved in
rotating flows. It provides improved pressure interpolation in situations where
large body forces or strong pressure variations are present as in swirling flows.
(b) Select Second Order Upwind from the adjacent drop-down list for Momentum,
Swirl Velocity, Turbulence Kinetic Energy, and Turbulence Dissipation Rate.
Use the scroll bar to access the turbulence discretization schemes.

!
(c) Retain the default Under-Relaxation Factors.
Note: For this problem, the default under-relaxation factors are satisfactory.
However, if the solution diverges or the residuals display large oscillations,
you may need to reduce the under-relaxation factors from their default
values.
See Section 26.10 of the User’s Guide for tips on how to adjust the under-
relaxation parameters for different situations.
7-12


(b) Click the OK button.
Note: For this calculation, the convergence tolerance on the continuity equation
is kept at 0.001. You can reduce this value if necessary, depending on the
behavior of the solution.

Write option, the history information will be lost when you exit FLUENT.

(c) Click Define... to specify the surface monitor parameters.

This will open the Define Surface Monitor panel.
i. Select Mass Flow Rate from the Report Type drop-down list.
ii. Select pressure-outlet-3 in the Surfaces list.
iii. Click OK to define the monitor.
(d) Click OK in the Surface Monitors panel to enable the monitor.
7-14

(b) Click Init and close the panel.
5. Save the case file (disk ke.cas).
Throughout the calculation, FLUENT will report reversed flow at the exit. This is
reasonable for the current case.
The solution should be sufficiently converged after about 230 iterations. The mass
flow rate history is shown in Figure 7.3.

0.0000
-0.0500
-0.1000
Mass -0.1500
Flow
Rate
-0.2000
-0.2500
-0.3000
0 25 50 75 100 125 150 175 200 225 250
Iteration
Convergence history of Mass Flow Rate on pressure-outlet-3

FLUENT 6.2 (axi, swirl, segregated, ske)
Figure 7.3: Mass Flow Rate History (k- Turbulence Model)

! verged, you should also check the net mass fluxes through the domain to
ensure that mass is being conserved.
7-16
(a) Select velocity-inlet-2 and pressure-outlet-3 under Boundaries.

(b) Keep the default Mass Flow Rate option.
(c) Click Compute.
!
continue iterating.
8. Save the data file (disk ke.dat).


Step 7: Postprocessing for the Standard k- Solution

1. Display the velocity vectors.
(a) Increase the Scale value to 50.

(b) Increase the Skip value to 1.
7-18
(c) Click Vector Options... to open the Vector Options panel.
i. Turn off the Z Component.

This allows you to examine the non-swirling components only.
ii. Click Apply and close the panel.
(d) Click Display in the Vectors panel to plot the velocity vectors.
A magnified view of the velocity field displaying a counter-clockwise circulation
of the flow is shown in Figure 7.4.
3.27e+00
3.10e+00
2.94e+00
2.78e+00
2.62e+00
2.45e+00
2.29e+00
2.13e+00
1.97e+00
1.80e+00
1.64e+00
1.48e+00
1.32e+00
1.15e+00
9.91e-01
8.29e-01
6.66e-01
5.04e-01
3.42e-01
1.79e-01
1.67e-02

Figure 7.4: Magnified View of Velocity Vectors within the Disk Cavity

2. Display filled contours of static pressure.

(a) Select Pressure... and Static Pressure in the Contours of drop-down list.
(b) Turn on the Filled option.
(c) Click Display.
The pressure contours are displayed in Figure 7.5. Notice the high pressure that
occurs on the right disk near the hub due to the stagnation of the flow entering from
the bore.
7-20
8.01e-01
7.34e-01
6.66e-01
5.98e-01
5.31e-01
4.63e-01
3.95e-01
3.28e-01
2.60e-01
1.92e-01
1.25e-01
5.70e-02
-1.07e-02
-7.84e-02
-1.46e-01
-2.14e-01
-2.81e-01
-3.49e-01
-4.17e-01
-4.84e-01
-5.52e-01

Figure 7.5: Contours of Static Pressure for Entire Disk Cavity
3. Create a constant y-coordinate line for postprocessing.

(a) Select Grid... and Y-Coordinate in the Surface of Constant drop-down lists.
(b) Click Compute to update the minimum and maximum values.
(c) Enter 37 in the Iso-Values field.
This is the radial position along which you will plot the radial velocity profile.

(d) Enter y=37cm for the New Surface Name.

(e) Click Create to create the isosurface.
Note: The name you use for an iso-surface can be any continuous string of
characters (without spaces).
4. Plot the radial velocity distribution on the surface y=37cm.

(a) Enable Node Values under Options.

(b) Select Velocity... and Radial Velocity from the Y Axis Function drop-down lists.
(c) Select the y-coordinate line y=37cm under Surfaces.
(d) Click Plot.
Figure 7.6 shows a plot of the radial velocity distribution along y = 37 cm.
(e) Save the radial velocity profile.
i. Select Write to File under Options.
ii. Click the Write... button.
iii. In the resulting Select File dialog box, specify ke-data.xy in the XY File
text entry box and click OK.
7-22
y=37
3.50e-01
3.00e-01
2.50e-01
2.00e-01
Radial 1.50e-01
Velocity
(m/s) 1.00e-01
5.00e-02
0.00e+00
-5.00e-02
0 1 2 3 4 5 6 7
Position (cm)
Radial Velocity
Figure 7.6: Radial Velocity Distribution: Standard k- Solution

Step 8: Solution Using the RNG k- Model

You will now recalculate the solution using the RNG k- turbulence model.
1. Turn on the RNG k- turbulence model with the enhanced near-wall treatment.
(a) Select RNG under k-epsilon Model.

(b) Enable the Differential Viscosity Model and Swirl Dominated Flow under RNG
Options.
The differential viscosity model and swirl modification can provide better ac-
curacy for swirling flows such as the disk cavity.
See Section 11.4.2 of the User’s Guide for more information.
(c) Keep the Enhanced Wall Treatment as the Near-Wall Treatment.
7-24
2. Continue the calculation by requesting 200 iterations.

The solution should converge after about 160 additional iterations.
3. Save the case and data files (disk rng.cas and disk rng.dat).

Step 9: Postprocessing for the RNG k- Solution

1. Plot the radial velocity distribution for the RNG solution and compare it with the
distribution for the standard k- solution.
(a) Load the k- data.

i. Click the Load File... button.
This will open the Select File dialog box.
ii. In the Select File dialog box, select the file ke-data.xy in the Files list.
iii. Click OK.
(b) In the Solution XY Plot panel, select Velocity... and Radial Velocity in the Y
Axis Function drop-down lists.
(c) In the Surfaces list, select y=37cm.
(d) Turn off the Write to File option.
7-26
(e) Click the Curves... button to define a different curve symbol for the RNG k-
data.
This will open the Curves - Solution XY Plot panel.
i. Set the Curve # to 0.

ii. Select x in the Symbol drop-down list.
iii. Click Apply and Close the panel.
(f) Click Plot in the Solution XY Plot panel.
y=37cm
y=37cm
4.00e-01
3.50e-01
3.00e-01
2.50e-01
2.00e-01
Radial 1.50e-01
Velocity
1.00e-01
(m/s)
5.00e-02
0.00e+00
-5.00e-02
-1.00e-01
0 1 2 3 4 5 6 7
Position (cm)
Radial Velocity Jan 06, 2005

FLUENT 6.2 (axi, swirl, segregated, rngke)
Figure 7.7: Radial Velocity Distribution: RNG and Standard k- Solutions
The plot should be similar to the one shown in Figure 7.7. The peak velocity
predicted by the RNG solution is higher than that predicted by the k- solution.
This is due to the less diffusive character of the RNG k- model.

(g) Adjust the range of the x axis to magnify the region of the peaks.
i. In the Solution XY Plot panel, click the Axes... button to specify the x-axis
range.
This will open the Axes - Solution XY panel.
ii. Deselect Auto Range under Options.

iii. Under Range, enter 0 for the Minimum and 1 for the Maximum.
iv. Click Apply and Close the panel.
v. Click Plot in the Solution XY Plot panel.
The difference between the peak values calculated by the two models is now
more apparent.
7-28
y=37cm
y=37cm
4.00e-01
3.50e-01
3.00e-01
2.50e-01
Radial 2.00e-01
Velocity
(m/s) 1.50e-01
1.00e-01
5.00e-02
0.00e+00
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Position (cm)
Radial Velocity
FLUENT 6.2 (axi, swirl, segregated, rngke)
Figure 7.8: Radial Velocity Distribution: RNG and Standard k- Solutions (x = 0 cm to
x = 1 cm)
Summary
This tutorial has demonstrated how to set up an axisymmetric disk cavity problem in
FLUENT. The ability to calculate a swirl velocity permits the use of a 2D mesh, so
the calculation is simpler and more economical to run. This is especially important
for problems where the enhanced wall treatment is used, and the near-wall flow field is
resolved through the viscous sublayer and buffer zones (that is, the first grid point away
from the wall is placed at a y + on the order of 1).
See Section 11.10 of the User’s Guide for more information about grid considerations for
turbulence modeling.

Further Improvements
The case modeled in this tutorial lends itself to parametric study due to its relatively
small size. Here are some things you may wish to try:
• Separate wall-6 into two walls.

Grid −→ Separate −→Faces...
Specify one wall to be stationary, and rerun the calculation.
• Use adaption to see if resolving the high velocity and pressure-gradient region of
the flow has a significant effect on the solution.
• Introduce a non-zero swirl at the inlet or use a velocity profile for fully-developed
pipe flow. This is probably more realistic than the constant axial velocity used
here, since the flow at the inlet is typically being supplied by a pipe.
• Model compressible flow (using the ideal gas law for density) rather than assuming
incompressible flow.
References
1. Pincombe, J.R., “Velocity Measurements in the Mk II - Rotating Cavity Rig with a
Radial Outflow”, Thermo-Fluid Mechanics Research Centre, University of Sussex,
Brighton, UK, 1981.
7-30
Tutorial 8. Using Multiple Rotating Reference Frames
Introduction
Many engineering problems involve rotating flow domains. One example is the centrifugal
blower unit that is typically used in automotive climate control systems. For problems
where all the moving parts (fan blades, hub and shaft surfaces, etc.) are rotating at a
prescribed angular velocity, and the stationary walls (e.g., shrouds, duct walls) are sur-
faces of revolution with respect to the axis of rotation, the entire domain can be referred
to as a single rotating frame of reference. However, when each of several parts is rotating
about a different axis of rotation, or about the same axis at different speeds, or when the
stationary walls are not surfaces of revolution (such as the volute around a centrifugal
blower wheel), a single rotating coordinate system is not sufficient to “immobilize” the
computational domain so as to predict a steady-state flow field.
In FLUENT, the flow features associated with multiple rotating parts can be analyzed
using the multiple reference frame (MRF) capability. This model is powerful in that
multiple rotating reference frames can be included in a single domain. The resulting flow
field is representative of a snapshot of the transient flow field in which the rotating parts
are moving. However, in many cases the interface can be chosen in such a way that the
flow field at this location is independent of the orientation of the moving parts. In other
words, if an interface can be drawn on which there is little or no angular dependence,
the model can be a reliable tool for simulating time-averaged flow fields. It is therefore
very useful in complicated situations where one or more rotating parts are present.
This tutorial illustrates the procedure for setting up and solving a problem using the
MRF capability. As an example, the flow field on a 2D section of a centrifugal blower
will be calculated. The example will be limited to a single rotating reference frame.
The following FLUENT features will be demonstrated in this tutorial:
• Specifying different frames of reference for different fluid zones.

• Setting the relative velocity of each wall.
• Calculating a solution using the segregated solver
Prerequisites

Using Multiple Rotating Reference Frames
In general, to solve problems using the MRF feature, you should be familiar with the
concept of creating multiple fluid zones in your grid generator.
Problem Description
This problem considers a 2D section of a generic centrifugal blower. A schematic of the
problem is shown in Figure 8.1. The blower consists of 32 blades, each with a chord length
of 13.5 mm. The blades are located approximately 56.5 mm (measured from the leading
edge) from the center of rotation. The radius of the outer wall varies logarithmically
from 80 mm to 146.5 mm. The total pressure at the inlet is defined to be 200 Pa and the
flow discharges to ambient conditions (static pressure = 0 Pa). The blades are rotating
with an angular velocity of 261 rad/s. The flow is assumed to be turbulent.
Pressure-inlet-5
261 rad/s
35 mm
56.5 mm
blower blades
(13.5 mm chord length)
Pressure-Outlet-9
145 mm
8-2
Setup and Solution

Preparation
1. Download multiple_rotating.zip from the Fluent Inc. User Services Center or
copy it from the FLUENT documentation CD to your working directory (as de-
scribed in Tutorial 1).
2. Unzip multiple_rotating.zip.
blower.msh can be found in the /multiple rotating folder created after unzipping
the file.
Step 1: Grid
1. Read in the mesh file (blower.msh).
2. Check the grid.

Grid −→Check
Note: FLUENT will perform various checks on the mesh and will report the progress
in the console window. Pay particular attention to the reported minimum
volume. Make sure this is a positive number.
3. Smooth and swap the grid.

Grid −→Smooth/Swap...
Node smoothing and face swapping will improve the mesh quality. This step is
recommended for triangular and tetrahedral meshes.

(a) Retain the default smoothing parameters and click Smooth.

(b) Click Swap repeatedly until the Number Swapped under Swap Info is zero.
4. Display the mesh (Figure 8.2).

The mesh consists of three fluid zones, fluid-13, fluid-14, and fluid-18. These are
reported in the console window when the grid is read. In the Grid Display panel, the
fluid zones are reported as interior zones interior-61, interior-62 and interior-66. In
a later step, you will learn how to associate a fluid zone with an interior zone. The
fluid zone containing the blades will be solved in a rotational reference frame.
The fluid zones are separated by wall boundaries. These boundaries were used in
the grid generator to separate the fluid zones, and will be converted to interior zones
when the boundary conditions are set later in this tutorial. Each of these wall zones
also has an associated “shadow wall” which was created by FLUENT when it read
the grid. Shadow walls are created whenever a wall has fluid zones on both sides.
8-4
Grid
Figure 8.2: Mesh of the 2D Centrifugal Blower

Step 2: Models
8-6


Step 3: Materials
You will use the default material, air, with its predefined properties, for all fluid zones.
No action is required in the panel.
Extra: If needed, you could modify the fluid properties for air or copy another material
from the database.
8-8

1. Change wall-2 and wall-3 type to interior.

The zones wall-2 and wall-3 are the interfaces between the three fluid zones. They
need to be changed to type interior, as discussed earlier. The resulting interior faces
are those that have fluid cells on both sides but do not require any boundary condi-
tions to be set.
(a) Select wall-2 in the Zone list and then select interior in the Type list.
FLUENT will prompt for confirmation before changing the zone type.
(b) Click Yes and FLUENT will fuse wall-2 and wall-2-shadow together to form
interior-2.

(c) Click OK to keep the default Zone Name.

(d) Repeat the previous steps to change wall-3 to an interior zone named interior-3.
2. Identify the rotating fluid zone (i.e., the zone containing the blades) by displaying
the mesh for each zone.
It is unclear when you read the grid which fluid zone corresponds to which interior
zone. While the interior zones can be selected individually in the Grid Display panel,
the fluid zones cannot. Commands in the text interface, however, can be used to
make this association.
(a) Deselect all surfaces by clicking on the unshaded icon to the right of Surfaces.
(b) Click the Outline button at the bottom of the panel to select only the outline
surfaces of the domain.
(c) Click Display.
Only the domain boundaries and interior walls will be displayed.
(d) In the console window, type the commands shown in boxes in the dialog below.
Hint: You may need to press the <Enter> key to get the > prompt.
> display
/display> zone-grid
()
zone id/name(1) [()] 13
zone id/name(2) [()] <Enter>
The resulting display (Figure 8.3) shows that zone fluid-13 corresponds to the ro-
tating region.
8-10
Thread Grid: (13)

Figure 8.3: Mesh in fluid-13
3. Define a rotational reference frame for fluid-13.

(a) Keep the Rotation-Axis Origin default setting of (0,0).
This is the center of curvature for the circular boundaries of the rotating zone.
(b) Select Moving Reference Frame from the Motion Type drop-down list.
Hint: Use the scroll bar to access the Motion Type list.
(c) Scroll down further, and set the Speed under Rotational Velocity to 261 rad/s.

Note: Since the other fluid zones are stationary, you do not need to set any bound-
ary conditions for them. If one of the remaining fluid zones was also rotating,
you would need to set the appropriate rotational speed for it.
4. Set the following conditions (see Figure 8.1) for the flow inlet (pressure-inlet-5).
8-12
Note: All pressures that you specify in FLUENT are gauge pressures, relative to
the operating pressure specified in the Operating Conditions panel. By default,
the operating pressure is 101325 Pa.
See Section 8.13 of the User’s Guide for details.
5. Set the backflow turbulence parameters for the flow outlet (pressure-outlet-9) to the
same values used for pressure-inlet-5.
Note: The backflow values are used only if reversed flow occurs at the outlet, but it
is a good idea to use reasonable values, even if you do not expect any backflow
to occur.
6. Define the velocity of the wall zone representing the blades (wall-7) relative to the
moving fluid zone.
With fluid-13 set to a rotating reference frame, wall-7 becomes a moving wall.
(a) In the Momentum section of the Wall panel, enable the Moving Wall option.
The panel will expand to show the wall motion parameters.
(b) Under Motion, select Relative to Adjacent Cell Zone and Rotational.
(c) Set the (relative) Speed to 0 rad/s.
The Rotation-Axis Origin should be located at x = 0 m and y = 0 m. With
these settings, the blades will move at the same speed as the surrounding fluid.

8-14
Step 5: Solution
1. Choose the second-order discretization scheme for the governing equations.
(a) In the drop-down lists next to Momentum, Turbulence Kinetic Energy, and
Turbulence Dissipation Rate, select Second Order Upwind.
The second-order scheme will provide a more accurate solution.
(b) Keep the default parameters for all other solution controls.


3. Initialize the solution using the boundary conditions set at pressure-inlet-5.

(a) Select pressure-inlet-5 in the Compute From drop-down list.

(b) Select Absolute under Reference Frame.
(c) Click Init to initialize the solution.
8-16
Note: In this tutorial, you chose an Absolute reference frame for initializing the
solution. In certain cases, Relative to Cell Zone may help the solution converge
faster.
See Chapter 26.15 of the User’s Guide for guidelines.
4. Save the case file (blower.cas).


During the calculation, FLUENT will report that there is reversed flow occurring at
the exit. This is due to the sudden expansion, which results in a recirculating flow
near the exit.
The solution will converge in around 160 iterations (when all residuals have dropped
below 0.001).
6. Save the case and data files (blower2.cas and blower2.dat).

Note: It is good practice to save the case file whenever you are saving the data.
This will ensure that the relevant parameters corresponding to the current
solution data are saved accordingly.

1. Display filled contours of total pressure (Figure 8.4).
(a) Select Pressure... and Total Pressure in the Contours of drop-down lists.
(c) Click Display.
Total pressure contours show the expected pressure jump across the blower blades.
8-18
1.13e+03
1.03e+03
9.25e+02
8.23e+02
7.22e+02
6.20e+02
5.18e+02
4.16e+02
3.15e+02
2.13e+02
1.11e+02
9.50e+00
-9.22e+01
-1.94e+02
-2.96e+02
-3.97e+02
-4.99e+02
-6.01e+02
-7.03e+02
-8.04e+02
-9.06e+02
Contours of Total Pressure (pascal)

Figure 8.4: Contours of Total Pressure


(a) Set the Scale factor to 5.

(b) Click Display to view the vectors.
By default, Auto Scale is chosen. This will automatically scale the length of velocity
vectors relative to the size of the smallest cell in the mesh. To increase the length
of the “scaled” vectors, set the Scale factor to a value greater than 1.
8-20
6.83e+01
6.49e+01
6.15e+01
5.81e+01
5.47e+01
5.13e+01
4.78e+01
4.44e+01
4.10e+01
3.76e+01
3.42e+01
3.08e+01
2.74e+01
2.39e+01
2.05e+01
1.71e+01
1.37e+01
1.03e+01
6.87e+00
3.45e+00
3.93e-02


3. Report the mass flux at pressure-inlet-5 and pressure-outlet-9.

(a) Keep the Mass Flow Rate setting under Options.

(b) Select pressure-inlet-5 and pressure-outlet-9 in the Boundaries list.
(c) Click Compute.
The net mass imbalance should be no more than a small fraction (say, 0.5%) of the
total flux through the system. If a significant imbalance occurs, you should decrease
your residual tolerances by at least an order of magnitude and continue iterating.
The flux report will compute fluxes only for boundary zones. To report fluxes on
surfaces or planes, use the Surface Integrals... option in the Report menu.
Summary
This tutorial illustrates the procedure for setting up and solving problems with multiple
reference frames using FLUENT. Although this tutorial considers only one rotating fluid
zone, extension to multiple rotating fluid zones is straightforward as long as you delineate
each fluid zone.
Note that this tutorial was solved using the default absolute velocity formulation. For
some problems involving rotating reference frames, you may wish to use the relative
velocity formulation. See the User’s Guide for details.
8-22
Tutorial 9. Using the Mixing Plane Model
Introduction
This tutorial considers the flow in an axial fan with a rotor in front and stators (vanes)
in the rear. This configuration is typical of a single-stage axial flow turbomachine. By
considering the rotor and stator together in a single calculation, you can determine the
interaction between these components.
• Use the standard k- model with standard wall functions
• Use a mixing plane to model the rotor-stator interface
• Compute and display circumferential averages of total pressure on a surface
Prerequisites
Problem Description
The problem to be considered is shown schematically in Figure 9.1. The rotor and stator
consist of 9 and 12 blades, respectively. A steady-state solution for this configuration
using only one rotor blade and one stator blade is desired. Since the periodic angles for
the rotor and stator are different, a mixing plane must be used at the interface.
The mixing plane is defined at the rotor outlet/stator inlet. The grid is set up with
periodic boundaries on either side of the rotor and stator blades. A pressure inlet is used
at the upstream boundary and a pressure outlet at the downstream boundary. Ambient
air is drawn into the fan (at 0 Pa gauge total pressure) and is exhausted back out to the
ambient environment (0 Pa static pressure). The hub and blade of the rotor are assumed
to be rotating at 1800 rpm.

Using the Mixing Plane Model
stator
rotor
outlet
inlet
z x ω = 1800 rpm
Setup and Solution

Preparation
1. Download mixing_plane.zip from the Fluent Inc. User Services Center or copy
Tutorial 1).
2. Unzip mixing_plane.zip.
fanstage.msh can be found in the /mixing plane folder created after unzipping
the file.
9-2
Step 1: Grid
1. Read the grid file (fanstage.msh).
2. Check the grid.

Grid −→Check
FLUENT will perform various checks on the grid and will report the progress in the
console window. Pay particular attention to the minimum volume. Make sure that
this is a positive number.

(a) Select rotor-blade, rotor-hub, rotor-inlet-hub, stator-blade, and stator-hub in the

Surfaces list.
(b) Click Display.
(c) Rotate the view to get the display shown in Figure 9.2.

Z X
Grid
Figure 9.2: Grid Display for the Multistage Fan
9-4
Step 2: Units
1. For convenience, define new units for angular velocity.
The angular velocity for this problem is known in rpm, which is not the default unit
for angular velocity. You will need to redefine the angular velocity units as rpm.
(a) Select angular-velocity under Quantities, and rpm under Units.


Step 3: Models
9-6
2. Turn on the standard k- turbulence model with standard wall functions.
(a) Under Model, select k-epsilon.

(b) Under k-epsilon Model, keep the default Standard option.
(c) Under Near-Wall Treatment, keep the default Standard Wall Functions option.

Step 4: Mixing Plane

In this step, you will create the mixing plane between the pressure outlet of the rotor and
the pressure inlet of the stator.
Define −→Mixing Planes...
1. Select pressure-outlet-rotor in the Upstream Zone list.
2. Select pressure-inlet-stator in the Downstream Zone list.
3. Click Create.
FLUENT will name the mixing plane by combining the names of the zones selected
as the Upstream Zone and Downstream Zone. This new name will be displayed in
the Mixing Plane list.
The essential idea behind the mixing plane concept is that each fluid zone (stator
and rotor) is solved as a steady-state problem. At some prescribed iteration inter-
val, the flow data at the mixing plane interface are averaged in the circumferential
direction on both the rotor outlet and the stator inlet boundaries. FLUENT uses
these circumferential averages to define “profiles” of flow properties. These profiles
are then used to update boundary conditions along the two zones of the mixing plane
interface.
In this example, profiles of averaged total pressure (p0 ), static pressure (ps ), direc-
tion cosines of the local flow angles in the radial, tangential, and axial directions
(αr , αt , αz ), total temperature (T0 ), turbulence kinetic energy (k), and turbulence
dissipation rate () are computed at the rotor exit and used to update boundary
conditions at the stator inlet. Likewise, the same profiles, except for that of total
9-8
pressure are computed at the stator inlet and used as a boundary condition on the
rotor exit. You can view the profiles computed at the rotor exit and stator inlet in
the Boundary Profiles panel.
Define −→Profiles...
You will also see that these profiles appear in the boundary conditions panels for
the rotor exit and stator inlet.
See Section 10.4 of the User’s Guide for more information on mixing planes.

Step 5: Materials
1. Accept the default properties for air.
For the present analysis, you will model air as an incompressible fluid with a density
of 1.225 kg/m3 and a dynamic viscosity of 1.7894× 10−5 kg/m-s. Since these are
the default values, no change is required in the materials panel.
9-10

1. Set the conditions for the rotor fluid (fluid-rotor).
(a) Under Rotation-Axis Direction, enter -1 next to Z .

According to the right-hand rule and Figure 9.1, the axis of rotation is the −Z
axis. You specify this by entering the vector (0, 0, −1) for the Rotation-Axis
Direction.
(b) Select Moving Reference Frame in the Motion Type drop-down list.
Hint: Use the scroll bar to access Motion Type.
(c) Set the Speed (under Rotational Velocity) to 1800 rpm.
Hint: Use the scroll bar to access Rotational Velocity.

2. Set the conditions for the stator fluid (fluid-stator).
(a) Under Rotation-Axis Direction, enter -1 next to Z.
3. Specify rotational periodicity for the periodic boundary of the rotor (periodic-11).
9-12
4. Specify rotational periodicity for the periodic boundary of the stator (periodic-22).
5. Set the following conditions for the pressure inlet of the rotor (pressure-inlet-rotor).
To model ambient conditions, you use P0 = 0 gauge. The turbulence level is as-
sumed to be low (1% ) and the hydraulic diameter is used as the length scale.

6. Examine the conditions for the pressure inlet of the stator (pressure-inlet-stator).
The profiles computed at the rotor outlet are used to update the boundary conditions
at the stator inlet. These profiles were set for you automatically when the mixing
plane was created. Therefore, you do not need to set any parameters in this panel.
7. Examine the conditions for the pressure outlet of the rotor (pressure-outlet-rotor).
The Backflow Direction Specification Method was set to Direction Vector when you
created the mixing plane, and the Coordinate System to Cylindrical (like for the stator
inlet ). The values for the direction cosines are taken from the profiles at the stator.
9-14
8. Set the conditions for the pressure outlet of the stator (pressure-outlet-stator).
(a) Select Radial Equilibrium Pressure Distribution.

Radial equilibrium is used to simulate the pressure distribution which exists

due to rotation according to
∂p ρv 2
= θ
∂r r
where vθ is the tangential velocity. This is a good approximation for axial flow
configurations with 0 straight flow paths (i.e., little change in radius from inlet
to exit).
(b) Retain the default Backflow Direction Specification Method.
In problems where a backflow exists at the pressure outlet boundary (e.g.,
torque-converter), you can use this option to specify the direction of the back-
flow.
(c) Select Intensity and Viscosity Ratio for the Turbulence Specification Method.
(d) Set the Backflow Turbulence Intensity to 1%.
(e) Set the Backflow Turbulent Viscosity Ratio to 1.
9. Set the conditions for the inlet hub of the rotor (rotor-inlet-hub).
9-16
(a) Select Moving Wall.

The panel will expand to show the wall motion inputs.
(b) Select Absolute and Rotational under Motion.
(c) Set the Rotation-Axis Direction by entering -1 next to Z.
These conditions set the rotor-inlet-hub to be a stationary wall in the absolute frame.
10. Set the conditions for the shroud of the rotor inlet (rotor-inlet-shroud).


These conditions set the rotor-inlet-shroud to be a stationary wall in the absolute
frame.
11. Set the following conditions for the rotor shroud (rotor-shroud).
9-18

These conditions set the rotor-shroud to be a stationary wall in the absolute frame.

12. Accept the default conditions for the rotor-hub.

For a rotating reference frame, FLUENT assumes by default that walls rotate with
the grid, and hence are moving with respect to the stationary (absolute) reference
frame. Since the rotor-hub is rotating, you should keep the default settings.
9-20
Step 7: Solution
(a) Under Discretization, select Second Order Upwind for Momentum.

(b) Select Power Law for Turbulence Kinetic Energy and Turbulence Dissipation Rate.
(c) Set the Under-Relaxation Factors for Pressure to 0.2, Momentum to 0.5, Tur-
bulence Kinetic Energy to 0.5, and Turbulence Dissipation Rate to 0.5.
Note: For this problem, it was found that these under-relaxation factors worked
well.
See Section 26.10 of the User’s Guide for tips on how to adjust the under-
relaxation parameters for different situations.



(b) Click OK.
9-22
(a) Increase the Surface Monitors value to 1.

write option, the history information will be lost when you exit FLUENT.
(c) Click on Define... to specify the surface monitor parameters in the Define
i. Select Mass Flow Rate from the Report Type drop-down list.
ii. Select pressure-outlet-stator in the Surfaces list.
iii. Click on OK to define the monitor.
(d) Click on OK in the Surface Monitors panel to enable the monitor.


(a) Select Absolute under Reference Frame.

For rotor-stator problems, initializing in the absolute frame is desirable, as
initializing in the relative frame would introduce a non-uniform swirl velocity
into the stationary domain.
(b) Set the initial value for Z Velocity to -1.
(c) Click on Init and close the panel.
5. Save the case file (fanstage.cas).


9-24
Calculating until the mass flow rate converges will require significant CPU
! resources. Instead of calculating the solution, you can read the data file
(fanstage.dat) with the pre-calculated solution, and proceed to the post-
processing section of the tutorial (Step 8). This data file can be found in
the directory where you found the mesh file.
The solution will converge after about 640 iterations. However, the residual history
plot is only one indication of solution convergence. Notice that the mass flow rate
has not reached a constant value. To remedy this, you will reduce the convergence
criterion for the continuity equation and iterate until the mass flow rate reaches a
constant value.
7. Save the case and data file (fanstage.cas & fanstage.dat).
8. Reduce the convergence criterion for the continuity equation.
(a) Set the Convergence Criterion for continuity to 1e-05 and click OK.
Note: In this case, the reason for continuing the calculation is to obtain bet-
ter global mass conservation; thus, only the convergence tolerance for the
continuity equation is adjusted. In general, the convergence behavior of
the continuity equation is a good indicator of the overall convergence of
the solution.

9. Request 1200 more iterations.

FLUENT will complete the given number of iterations. But since after a total of
about 1400 iterations the mass flow rate has leveled off, we can consider that the
solution is converged. The mass flow rate history is shown in Figure 9.3.
Monitors
monitor-1
-0.0020
-0.0040
-0.0060
-0.0080
-0.0100
Mass -0.0120
Flow
-0.0140
Rate
(kg/s) -0.0160
-0.0180
-0.0200
-0.0220
0 200 400 600 800 1000 1200 1400 1600 1800 2000
Y
Z X
Iteration
Convergence history of Mass Flow Rate on pressure-outlet-stator

Figure 9.3: Mass Flow Rate History
10. Save the case and data file (fanstage1.cas & fanstage1.dat).

! verged, you should also check the mass fluxes through the domain to ensure
that mass is being conserved.
9-26
(a) Select pressure-outlet-stator, pressure-inlet-stator, pressure-inlet-rotor, and pressure-

outlet-rotor under Boundaries.
(b) Keep the default Mass Flow Rate option and click on Compute.
!
continue iterating.
Note: The fluxes for the portions of the rotor and stator that have been modeled are
different. However, the flux for the whole rotor and the whole stator are very
nearly equal: approximately 0.23274 kg/s (0.02586 × 9 rotor blades), versus
approximately 0.23328 kg/s (0.01944 × 12 stator blades).

1. Create two surfaces for postprocessing, one at y = 0.12 m and one at z = −0.1 m.
The surface y = 0.12 m is a midspan slice through the grid. This view is good
for looking at the blade-to-blade flow field. The surface z = −0.1 m is an axial
plane downstream of the stator. This will be used to plot circumferentially-averaged
profiles.
(a) Select Grid... and Y-Coordinate in the Surface of Constant lists.

(b) Click on Compute to update the minimum and maximum values.
(c) Enter 0.12 in the Iso-Values field.
(d) Enter y=0.12 for the New Surface Name.
(e) Click on Create to create the isosurface.
(f) Select Grid... and Z-Coordinate in the Surface of Constant lists.
(g) Click on Compute to update the minimum and maximum values.
(h) Enter -0.1 in the Iso-Values field.
(i) Enter z=-0.1 for the New Surface Name.
Note: The default name that FLUENT gave the surface, z-coordinate-17, in-
dicates that this is surface number 17. This fact will be used later in the
tutorial when you plot circumferential averages.
(j) Click on Create to create the isosurface.
9-28
2. Display velocity vectors on the midspan surface y = 0.12.

(a) Select y=0.12 in the Surfaces list.

(b) Increase the Scale value to 10.
(c) Increase the Skip value to 2.
(d) Select arrow from the Style drop-down list.
(e) Click on Display to plot the velocity vectors.
(f) Rotate and zoom the view to get the display shown in Figure 9.4.

2.77e+01
2.63e+01
2.49e+01
2.35e+01
2.22e+01
2.08e+01
1.94e+01
1.80e+01
1.66e+01
1.53e+01
1.39e+01
1.25e+01
1.11e+01
9.73e+00
8.35e+00
6.97e+00
5.58e+00
4.20e+00
2.82e+00
1.44e+00 X
Y
5.63e-02
Z

Figure 9.4: Velocity Vectors on y = 0.12 Near the Stator Blade
Plotting the velocity field in this manner gives a good indication of the midspan
flow over the stator. For the rotor, it is instructive to similarly plot the relative
velocity field.
3. Plot a circumferential average of the total pressure on the plane z = −0.1.

(a) In the console window, type the commands shown in boxes in the dialog below.
Note: Surface 17 is the surface z = −0.1 you created earlier. For increased
resolution, 15 bands are used instead of the default 5.
> plot
/plot> circum-avg-radial
averages of> total-pressure
on surface [] 17
number of bands [5] 15
9-30
(b) Enter the name of the output file as circum-plot.xy when prompted.
Computing r-coordinate ...

Clipping to r-coordinate ... done.
Computing "total-pressure" ...
Computing averages ... done.
Creating radial-bands surface (32 31 30 29 28 27 26 25 24 23 22 21 20 19 18).
filename [""] "circum-plot.xy"
order points? [no]
(c) Display the circumferential average.

Plot −→File...
i. Click Add... and select the file circum-plot.xy.

ii. Click Plot.
The radial variation in the total pressure can be seen to be very non-uniform
in this plot (Figure 9.5). This implies that losses are largest near the hub.

3.75e+01
3.50e+01
3.25e+01
3.00e+01
2.75e+01
Total
Pressure 2.50e+01
2.25e+01
2.00e+01
1.75e+01
1.50e+01
0.10 0.11 0.11 0.12 0.12 0.12 0.13 0.14 0.14
Y
Z X Radius
Circumferential Averages
Figure 9.5: Plot of Circumferential Average of the Total Pressure on the Plane z = −0.1.
4. Display filled contours of total pressure.

9-32
(a) Select rotor-blade and rotor-hub in the Surfaces list.

(b) Select Pressure... and Total Pressure in the Contours of drop-down lists.
(c) Turn on the Filled option.
(d) Click Display.
The pressure contours are displayed in Figure 9.6. Notice the high pressure that
occurs on the leading edge of the rotor blade due to the motion of the blade.
5.04e+02
4.66e+02
4.29e+02
3.91e+02
3.53e+02
3.15e+02
2.77e+02
2.39e+02
2.02e+02
1.64e+02
1.26e+02
8.82e+01
5.04e+01
1.26e+01
-2.52e+01
-6.30e+01
-1.01e+02
-1.39e+02
-1.76e+02 Y
-2.14e+02
Z X
-2.52e+02
Contours of Total Pressure (pascal)

Figure 9.6: Contours of Total Pressure for the Rotor Blade and Hub
Summary
This tutorial has demonstrated the use of the mixing plane model for a typical axial flow
turbomachine configuration. The mixing plane model is useful for predicting steady-
state flow in a turbomachine stage, where local interaction effects (such as wake and
shock wave interaction) are secondary. If local effects are important, then an unsteady,
sliding mesh calculation is required.

9-34
Tutorial 10. Using Sliding Meshes
Introduction
In this tutorial, the sliding mesh capability of FLUENT is used to predict the time-
dependent flow through a two-dimensional rotor-stator blade row. The time-varying
rotor-stator interaction is modeled by allowing the mesh associated with the moving
rotor to translate (slide) relative to the stationary mesh associated with the stator blade.
• Merge two meshes into a single mesh, using tmerge

• Define boundary conditions and create grid-interface planes for sliding mesh simu-
lations
• Calculate a steady-state solution (using the coupled explicit solver) as an initial
guess for a transient flow prediction
• Calculate a transient solution using the second-order implicit unsteady formulation
and the coupled explicit solver
• Monitor solution history for time-dependent parameters
• Postprocess and store transient data sets, using the automatic procedures available
in FLUENT
Prerequisites
Problem Description
The rotor-stator geometry considered in this tutorial is shown in Figure 10.1. The geom-
etry consists of a planar slice through the rotor and stator blades, extracted by unrolling
a plane of constant radius (R = 0.686 m) in an axial flow turbomachine. The speed
of rotation, 410 RPM, yields a linear velocity of the rotor, RΩ, equal to 29.4 m/s, as
indicated in the figure. The fluid, assumed to be air, enters the stator row at the specified
total pressure and temperature and exits the rotor at the specified exit static pressure.
The inlet Mach number is 0.07 and the flow will be treated as compressible.

Using Sliding Meshes
Stator Vanes Rotor Blades

(stationary) (moving)
Cx
= 0.1524 m
P01 = 101325 Pa
blade pitch = 0.1959 m
T01 = 300 K
M = 0.07
P2 = 97576 Pa
sliding interface direction of

motion
(V = 29.445 m/s)
Figure 10.1: Rotor-Stator Problem Description
10-2
Setup and Solution

Preparation
1. Download sliding_mesh.zip from the Fluent Inc. User Services Center or copy
Tutorial 1).
2. Unzip sliding_mesh.zip.
rotor.msh and stator.msh can be found in the /sliding mesh folder created after
unzipping the file.
Note: The geometries of the rotor and stator flow domains have been meshed separately.
This is the usual procedure when the sliding mesh capability is used: separate mesh
files are created for the sliding and stationary mesh regions. This ensures that the
sliding interface between the two regions is defined by two separate boundary zones
which share no common nodes. The two separate mesh files must be merged prior
to reading them into FLUENT, as detailed in Step 1, below.

Step 1: Merging the Mesh Files

1. Start tmerge by typing utility tmerge -2d at the system prompt.
2. Provide the mesh file names, rotor.msh and stator.msh, as prompted. Provide
scaling factors of 1 and translations and rotations of zero for each mesh file. Save
the new merged mesh file as slide.msh.
Append 2D grid files.

tmerge2D Fluent Inc, Version 2.1.11
Enter name of grid file (ENTER to continue) : rotor.msh
x,y scaling factor, eg. 1 1 : 1 1
x,y translation, eg. 0 1 : 0 0
rotation angle (deg), eg. 45 : 0
Enter name of grid file (ENTER to continue) : stator.msh
x,y scaling factor, eg. 1 1 : 1 1
x,y translation, eg. 0 1 : 0 0
rotation angle (deg), eg. 45 : 0
Enter name of grid file (ENTER to continue) : <ENTER>
Enter name of output file : slide.msh
The mesh files must be read into tmerge in this order for the tutorial to run
! as written. Otherwise, zone names and numbers will be assigned differently
when the files are merged together. In general, however, you can specify
files to be read into tmerge in any order.
10-4
Step 2: Grid
2. Read in the mesh file slide.msh.

3. Check the grid.

Grid −→Check
4. Scale the grid.

Grid −→Scale...
(a) Select in in the Units Conversion drop-down list to complete the phrase Grid
Was Created In in (inches).
(b) Click on Scale to scale the grid.
The final domain extents should appear as in the panel above.


Note: You can use the mouse probe button (right button, by default) to find out the
boundary zone labels. As annotated in Figure 10.3, the upstream boundary is a
pressure inlet, the downstream boundary is a pressure outlet, and the lateral top
and bottom boundaries are periodic. The stator blade and stator-side fluid are
identified as wall-7 and fluid-9. The rotor blade and rotor-side fluid are wall-16
and fluid-18. (Your mouse probe will report the fluid regions as interior-8 and
interior-17 zones, for the stator and rotor sides, respectively. These are the
face zones associated with the fluid regions.)
To determine which fluid zone is the stator-side fluid, you can create the fluid-
9 and fluid-18 display surfaces using the Zone Surface panel. Then, display the
zones (one at a time) using the Grid Display panel.
If you wish to annotate your own graphics display, you can use the Annotate
panel.
Display −→Annotate...
10-6
Grid
Figure 10.2: Rotor-Stator Mesh Display
periodic-8
periodic-15
fluid-18
pressure-
pressure-outlet-14
inlet-3 wall-7
wall-16
Grid
Figure 10.3: Annotated Mesh

Step 3: Models
1. Select the coupled explicit solver.
Note: Initially, you will solve for the steady flow through the blade passage. Later,
after obtaining the steady flow as the starting point for the transient calcula-
tion, you will revisit this panel to turn on time-dependent flow.
10-8
2. Enable the standard k- turbulence model.

Note: The Reynolds number of the flow is about 105 , and the flow will be treated
as fully turbulent.

Step 4: Materials
1. Select air (the default material) as the fluid material, and use the ideal-gas law to
compute density. Retain the default values for all other properties.
Do not forget to click the Change/Create button after selecting ideal-gas in

!
the drop-down list for Density.
10-10

1. Set the operating pressure to 0 Pa.
Here, the operating pressure is set to zero and boundary condition inputs for pressure
will be defined in terms of absolute pressures. Boundary condition inputs should
always be relative to the value used for operating pressure.


1. Set the conditions for the upstream boundary (pressure-inlet-3).
(a) Change the Zone Name from pressure-inlet-3 to pressure-inlet.

(b) Set the Gauge Total Pressure to 101325 Pa.
(c) Set the Supersonic/Initial Gauge Pressure to 100978.2 Pa.
The inlet static pressure estimate is computed from the assumed inlet total
pressure and Mach number (see Figure 10.1).
(d) Set the Total Temperature to 300 K.
(e) In the Direction Specification Method drop-down list, select Direction Vector.
(f) In the Turbulence Specification Method drop-down list, select Intensity and Hy-
draulic Diameter.
(g) Set the Turbulence Intensity to 5%, and the Hydraulic Diameter to 0.1959 m.
The inlet turbulence length scale will be computed using the blade pitch as an
equivalent “hydraulic diameter”.
10-12
2. Set the conditions for the exit plane boundary (pressure-outlet-14).
(a) Change the Zone Name from pressure-outlet-14 to pressure-outlet.

(b) Set the Gauge Pressure to 97576 Pa.
(c) Set the Backflow Total Temperature to 300 K.
(d) In the Turbulence Specification Method drop-down list, select Intensity and Hy-
draulic Diameter.
(e) Set the Turbulence Intensity to 5%, and the Hydraulic Diameter to 0.1959 m.
Note: The temperature and turbulence conditions you input at the pressure outlet
will be used only if flow enters the domain through this boundary. You can set
them equal to the inlet values, as no flow reversal is expected at the domain exit
in this problem. In general, however, it is important to set reasonable values
for these downstream scalar values, in case flow reversal occurs at some point
during the calculation.

3. Keep the default Momentum boundary conditions for the stator blades (wall-7) and
the rotor blades (wall-16).
The velocity of a “non-moving” wall is assumed to match that of the adjacent mesh
region, yielding a no-slip condition in the reference frame of the mesh. Thus, FLU-
ENT will assume that the stator blade is stationary in the non-moving reference
frame of the stator mesh. Similarly, FLUENT will assume that the rotor blade is
moving at the grid speed in the sliding rotor region. Therefore, you will not modify
the wall velocity of the rotor (wall-16), even though the rotor is moving. The default
setting of a non-moving wall is correct and implies zero velocity in the moving ref-
erence frame of the sliding region. (The motion of the mesh region will be defined
as a boundary condition for the fluid zone, below.)
10-14
4. Set the conditions for the stator-side fluid (fluid-9).
(a) Change the Zone Name from fluid-9 to fluid-stator.

(b) Keep the default selection of air as the Material Name, and the Motion Type as
Stationary.
Hint: Use the scroll bar to access the Motion Type field.

5. Set the conditions for the rotor-side fluid (fluid-18).
(a) Change the Zone Name from fluid-18 to fluid-rotor.

(b) Keep the default selection of air as the Material Name, and the Motion Type as
Stationary.
Later, after solving the steady flow through the non-moving rotor passage, you will
return to this panel to specify that the rotor zone is sliding, as illustrated in Fig-
ure 10.1.
10-16
6. Define the zones on the sliding boundary as interface zones.

The sliding grid interface contains two boundary zones: pressure-inlet-12 and pressure-
outlet-5. pressure-inlet-12 is the upstream boundary of the rotor-side fluid region,
and pressure-outlet-5 is the downstream boundary of the stator-side fluid region.
These boundaries were defined as the upstream and downstream boundaries of the
original mesh files, rotor.msh and stator.msh, and must be redefined as interface
boundary types for the merged mesh.
(a) Select pressure-inlet-12 in the Zone list and choose interface as the new Type.
(b) Confirm that it is OK to change the boundary type.
(c) Change the Zone Name to interface-rotor.
(d) Repeat this procedure to convert pressure-outlet-5 to an interface boundary

named interface-stator.


In this step, you will create a periodic grid interface between the rotor and stator mesh
regions.
1. Select interface-rotor in the Interface Zone 1 list.

Note: In general, when one interface zone is smaller than the other, it is recom-
mended that you choose the smaller zone as Interface Zone 1. In this case,
however, since both zones are the same size, the order is not significant.
2. Select interface-stator in the Interface Zone 2 list.
3. Enter the name interface-rotor-stator under Grid Interface.
4. Select Periodic under Interface Type, and click Create.

The interface between the sliding and non-sliding zones will be treated as periodic
where the two zones are non-overlapping.
10-18
Step 8: Solution: Steady Flow with Non-Moving Rotor

1. Initialize the solution for steady flow.
(a) Select pressure-inlet in the Compute From drop-down list.



(a) Set the Courant Number to 2.

Setting the Courant number to 2.0 promotes rapid convergence for the steady
flow simulation. (For more information on the Courant number, see Step 9:
Enable Time Dependence and Sliding Rotor Motion).
(b) Change Multigrid Levels to 5.
Five levels of multigrid enable boundary condition information to propagate
rapidly across the solution domain.
(c) Under Discretization, select Second Order Upwind for Turbulence Kinetic Energy
and Turbulence Dissipation Rate.
Second-order discretization provides optimum accuracy.
10-20


4. Enable the monitoring of the lift force on the rotor blade (wall-16).
Note: Monitoring forces on the rotor blade provides a good measure of convergence
during the initial steady-state flow prediction. Here, you will request dynamic
plotting of lift as the solution proceeds. In addition, you will write the lift
information to a file, cl-hist.ss. You could choose to monitor any other
variable (e.g., mass flow rate), including a custom field function.
(a) Select Lift in the Coefficient drop-down list.

(b) In the Wall Zones list, select wall-16 (the rotor).
(c) Enable the Plot and Write options.
(d) In the File Name field, enter cl-hist.ss.
(e) Keep the Plot Window set to 1.
(f) Click Apply, and Close the panel.
10-22
5. Set the reference values to be used in the lift coefficient calculation.

Report −→Reference Values...
(a) Select pressure-inlet in the Compute From drop-down list.

(b) Change the Area to 0.1524 m2 .
(c) Change the Length to 0.1524 m (the axial chord length).
6. Save the steady flow case file (slide ss.cas).
The residual history and lift force history will be displayed as the calculation pro-
ceeds. The lift history should be similar to Figure 10.4.
Note: After 500 iterations, the steady flow calculation may not be fully converged.
Here, this is not of concern, as the steady-state prediction will be used only as
a starting solution for the transient sliding-mesh calculation.

0.0000
-2.0000
-4.0000
-6.0000
Cl -8.0000
-10.0000
-12.0000
-14.0000
-16.0000
0 50 100 150 200 250 300 350 400 450 500
Iterations
Lift Convergence History

FLUENT 6.2 (2d, coupled exp, ske)
Figure 10.4: Lift Coefficient History: Steady Flow, Non-Moving Rotor
8. Save the case and data files (slide ss.cas and slide ss.dat).
Note: If you choose a file name that already exists in the current directory, FLU-
ENT will prompt you for confirmation to overwrite the file.
10-24
9. Display the steady flow velocity vectors.

(a) Change the Scale to 10.

(b) Click Display.
The steady flow prediction shows the expected form, with peak velocity of about 140
m/s through the passage.

10. Enable the display of three periodic repeats of the solution domain (Figure 10.5).
(a) Under Periodic Repeats, click the Define... button.

This will open the Graphics Periodicity panel.
(b) Set the number of Number of Repeats to 4.

(c) Click Set.
This will reset Y Translation to the value shown above.
(d) Close the Graphics Periodicity panel.
(e) In the Views panel, click Apply.
The grid display will be updated to show three periodic repeats. You may need to
translate the display using your mouse to get the view shown in Figure 10.5.
10-26
1.42e+02
1.35e+02
1.28e+02
1.20e+02
1.13e+02
1.06e+02
9.93e+01
9.23e+01
8.53e+01
7.82e+01
7.12e+01
6.41e+01
5.71e+01
5.00e+01
4.30e+01
3.60e+01
2.89e+01
2.19e+01
1.48e+01
7.79e+00
7.49e-01

Figure 10.5: Velocity Vectors: Steady Flow, Non-Moving Rotor

11. Display the steady flow contours of static pressure (Figure 10.6).
The steady flow prediction shows the expected pressure distribution through the pas-
sage, with low pressure on the suction surfaces and stagnation around the impinge-
ment point on the rotor.
10-28
1.02e+05
1.01e+05
1.01e+05
9.99e+04
9.92e+04
9.86e+04
9.79e+04
9.72e+04
9.65e+04
9.59e+04
9.52e+04
9.45e+04
9.38e+04
9.31e+04
9.25e+04
9.18e+04
9.11e+04
9.04e+04
8.97e+04
8.91e+04
8.84e+04

Figure 10.6: Contours of Static Pressure: Steady Flow, Non-Moving Rotor

Step 9: Enable Time Dependence and Sliding Rotor Motion

In this step you will enable the rotor motion by turning on time dependence and setting
the sliding velocity of the rotor fluid zone.
1. Enable a time-dependent flow calculation.


(b) Under Unsteady Formulation, select 2nd-Order Implicit.
Implicit (dual) time-stepping allows you to set the physical time step used for the
transient flow prediction (while FLUENT continues to determine the time step used
for inner iterations based on a Courant condition). Here, second-order implicit
time-stepping is enabled: this provides higher accuracy in time than the first-order
option.
In explicit (global) time-stepping, FLUENT uses a single time step for the transient
calculation. This time step is based on the Courant condition:
CFL∆x
∆t =
λmax
10-30
where CFL is the Courant number. Explicit time-stepping might be the optimum
solution strategy if you are modeling a traveling shock wave, where the Courant
condition is ideal for determination of the time step value.
2. Define the sliding motion of the rotor-side fluid zone (fluid-rotor).

(a) Scroll down and select Moving Mesh in the Motion Type drop-down list.
(b) Change the Translational Velocity in the Y direction to -29.445.
3. Save the case and data files (slide un.cas and slide un.dat).
The mesh changes during the preview so be sure to save the case before mesh pre-
view.

4. Preview the sliding motion of the rotor-side fluid zone.

Solve −→Mesh Motion...
(a) Set the Time Step Size to 1.4e-4.

(b) Under Number of Time Steps, enter 50.
(c) Click Preview.
The graphics display will preview the sliding motion of the rotor-stator grid
geometry.
5. Read the case and data files back into FLUENT (slide un.cas and slide un.dat).
Note: As the mesh preview option advances the time step, thereby updating the
mesh, it is essential to reread the case file to ensure that the mesh is at the
proper position before the calculation is started. If the case file is not read
again and the solution is initialized after the mesh preview, the solution time
is initialized to zero but the mesh does not go back to its original position.
10-32
Step 10: Solution: Unsteady Flow with Moving Rotor

1. Change the Courant number to 1 for the transient calculation.
The Courant number controls the time step used by FLUENT during the inner iter-
ations performed during each time step. A Courant number of 1 is a conservative
setting that should ensure the stability of the inner iterations.

2. Reset the lift force monitor.

(a) In the File Name field, enter cl-hist.td.

(b) Click Apply.
(c) Click Clear.
This will remove the lift coefficient data for the steady-state calculation (i.e.,
the contents of the file cl-hist.ss) from memory.
(d) Click Yes when asked to confirm the data discard.
If you do not clear the old force-monitoring data, FLUENT will plot them
! with the new data, corrupting the new lift coefficient plot.
(e) Click on the Axes... button.
This will open the Axes - Force Monitor Plot panel.
10-34
(f) Under Axis, select X (the default).

(g) Under Number Format, check that the Type is float.
(h) Set the Precision to 4.
(i) Click Apply.
(j) Under Axis, select Y.

(k) Under Number Format, check that the Type is float and the Precision is 4.
(l) Close the Axes - Force Monitor Plot and Force Monitors panels.

Note: Monitoring the lift force is an ideal way to determine when the transient flow
prediction becomes time-periodic (independent of the initial condition). In the
time-periodic solution, the lift force variation will repeat identically from one
passing period to the next.
3. Set the time step parameters.

(a) Set the Time Step Size to 0.0001 second.

(b) Check that the Max Iterations per Time Step is set to 20.
(c) Click Apply.
Note: The selection of the time step is critical for accurate time-dependent flow
predictions. Here, the time step is chosen to be about 1/70 of the passing
period, T. The passing period is the time it takes for the rotor blade to pass
from one stator blade row to the next:
T = (0.1959 m)/(29.4 m/s) = 6.7 × 10−3 sec
Using a time step of 0.0001 second, 67 time steps will be performed as the
rotor performs one pass.
10-36
The maximum number of iterations per time step should be set large
! enough so that the inner iterations converge before the solution moves
to the next time value. The value of 20, selected here, is quite small: it
is likely that the initial time steps will not fully converge within 20 inner
iterations. While this would be unsuitable for prediction of most time-
dependent flows, the current simulation does not require high accuracy
during the initial time steps. The rotor-stator flow prediction will be con-
tinued in time until a time-periodic flow is obtained. Low accuracy during
the initial passing periods is acceptable as long as convergence is achieved
during each time step of the final passing periods.
4. Save the transient sliding mesh case file (slide td.cas).

5. Start the transient calculation by requesting 1000 time steps.

Calculation of 1000 time steps will require significant CPU resources. In-
! stead of calculating, you can read the case and data files saved after 0.1
seconds:
slide01.cas and slide01.dat
After reading the files, skip to Step 11: Postprocessing at t=0.1 Second.
(The case and data files are available in the same directory where you found
the mesh files.)
By requesting 1000 time steps, you are asking FLUENT to compute until time is
equal to 0.1 second. This will include roughly 15 passing periods (15 × 6.7e-3 sec
= 0.10 sec). Experience shows that the flow becomes time-periodic after about 12
passing periods. The lift history display allows you to confirm this. Your lift force
history should be similar to that shown in Figure 10.7. Notice the periodicity of the
lift coefficient after approximately 0.05 seconds.
1.00e+01
5.00e+00
0.00e+00
-5.00e+00
Cl
-1.00e+01
-1.50e+01
-2.00e+01
-2.50e+01
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
Time
Lift Convergence (Time=1.0000e-01)

FLUENT 6.2 (2d, coupled exp, ske, unsteady)
Figure 10.7: Lift Coefficient History: Unsteady Flow, Moving Rotor
6. Reset the lift coefficient history range to focus on the periodicity.

(a) Click the Axes... button.
10-38
i. Under Axis, select X.

ii. Under Options, deselect Auto Range.
iii. Under Range, set the Minimum to 0.02, and the Maximum to 0.1.
iv. Click Apply.
v. Under Axis, select Y.

vi. Under Options, deselect Auto Range.
vii. Under Range, set the Minimum to -9.2, and the Maximum to -7.6.

viii. Click Apply and Close the panel.

(b) In the Force Monitors panel, click Plot.
In Figure 10.8, you can see the periodicity more clearly.
-7.4000
-7.6000
-7.8000
-8.0000
-8.2000
Cl
-8.4000
-8.6000
-8.8000
-9.0000
-9.2000
0.020 0.030 0.040 0.050 0.060 0.070 0.080 0.090 0.100 0.110
Time

Figure 10.8: Lift Coefficient History: Narrowed Range
7. Save the case and data files at t = 0.1 second (slide01.cas and slide01.dat).
When the sliding mesh model is used, you must save a case file whenever a
! data file is saved. This is because the case file contains the grid information,
which is changing with time.
10-40
Step 11: Postprocessing at t = 0.1 Second

The solution data saved at t = 0.1 second can be reviewed using any of FLUENT’s post-
processing features. Time-dependent data are analyzed just like steady-state results: you
read a case and data file for each time value of interest. Note that this means you must
save case and data files (or graphics files) at all intermediate time values for which results
are of interest. Automatic saving of results during a transient calculation is used in Step
12, Saving and Postprocessing Time-Dependent Data Sets.
For sliding mesh cases, it is important that you read the associated case
! file whenever you read a data file, because the case file contains the grid
information, which is changing with time.
1. Display the velocity vectors at t = 0.1 second (Figure 10.9).

9.27e+01
8.82e+01
8.36e+01
7.91e+01
7.45e+01
6.99e+01
6.54e+01
6.08e+01
5.63e+01
5.17e+01
4.72e+01
4.26e+01
3.80e+01
3.35e+01
2.89e+01
2.44e+01
1.98e+01
1.52e+01
1.07e+01
6.13e+00
1.57e+00

Figure 10.9: Velocity Vectors at T = 0.1 Second: Unsteady Flow
Note: The velocity vectors in Figure 10.9 are displayed with respect to the absolute
reference frame (default). If you want to display the vectors with respect to the
moving reference frame, you can select Relative Velocity in the Vectors of drop-
down list and select the rotor fluid zone as the Reference Zone in the Reference
Values panel.

2. Display contours of static pressure at t = 0.1 second (Figure 10.10).

1.01e+05
1.01e+05
1.01e+05
1.00e+05
9.98e+04
9.94e+04
9.90e+04
9.86e+04
9.82e+04
9.78e+04
9.75e+04
9.71e+04
9.67e+04
9.63e+04
9.59e+04
9.55e+04
9.51e+04
9.47e+04
9.43e+04
9.39e+04
9.35e+04
Contours of Static Pressure (pascal) (Time=1.0000e-01)

Figure 10.10: Contours of Static Pressure at Time = 0.1 Second: Unsteady Flow
Note: Slight discontinuities in the pressure contours along the sliding interface are
expected. This is because the contour plotting uses one-sided interpolation on
either side of the sliding plane. This is purely a display issue.
10-42
3. Determine the instantaneous total pressure loss through the system.

Report −→Surface Integrals...
(a) In the Report Type drop-down list, select Mass-Weighted Average.

(b) Select Pressure... and Total Pressure in the Field Variable drop-down lists.
(c) In the Surfaces list, select pressure-outlet.
(d) Click Compute.
The mass-averaged (instantaneous) total pressure at the exit is about 98100
Pa, implying a loss of about 3200 Pa from the inlet total pressure of 101325
Pa.
4. Plot the instantaneous pressure coefficient on the rotor blade at t = 0.1 second
(Figure 10.11).

(a) Select Pressure... and Pressure Coefficient in the Y Axis Function drop-down
lists.
(b) In the Surfaces list, select wall-16.
(c) Click Plot.
wall-16
-2.00e+00
-4.00e+00
-6.00e+00
-8.00e+00
-1.00e+01
Pressure -1.20e+01
Coefficient
-1.40e+01
-1.60e+01
-1.80e+01
-2.00e+01
-2.20e+01
0.175 0.2 0.225 0.25 0.275 0.3 0.325 0.35 0.375 0.4
Position (m)
Pressure Coefficient (Time=1.0000e-01)

Figure 10.11: Pressure Coefficient on the Moving Rotor at Time = 0.1 Second
10-44
Step 12: Saving and Postprocessing Time-Dependent Data Sets

After 0.1 second, the sliding mesh flow prediction has achieved a time-periodic state. In
order to study how the flow changes within a single passing period, you will now continue
the time-marching through the next period and save results every 5 time steps.
1. Request saving of case and data files every 5 time steps.

(b) In the Filename field, enter slide.gz.
(c) Click OK.
FLUENT will append the time step value to the file name prefix (slid). The
standard extensions (.cas and .dat will also be appended. This will yield file
names of the form slid1005.cas and slid1005.dat, where 1005 is the time
step number.
The extension “.gz”, supplied in the panel above, instructs FLUENT to save
the case and data files in compressed format, yielding file names of the form
slid1005.cas.gz and slid1005.dat.gz. If you leave off the “.gz” exten-
sion, FLUENT will save the files as slid1005.cas and slid1005.dat, etc.
When the sliding mesh model is used, you must save a case file whenever a
! data file is saved. This is because the case file contains the grid information,
which is changing with time.
Extra: If you want to generate a solution animation by plotting, for example, pres-
sure contours every 5 time steps, you can use the Solution Animation panel to
set up the animation before you begin the calculation. Tutorial 4 demonstrates
how to do this.

2. Reset the lift force monitor.

Supplying a new file name eliminates overwriting of the file stored during the pre-
vious time steps.
(a) In the File Name field, enter cl-hist.cyc.

(b) Click Apply.
(c) Click Clear.
This will remove the lift coefficient data for the previous time steps from mem-
ory (i.e., cl-hist.td).
(d) Click Yes when asked to confirm the data discard.
(e) Click on the Axes... button to reset the domain and range.
i. Under Axis, select X.
ii. Under Options, select Auto Range.
This will deactivate the Minimum and Maximum range fields.
iii. Click Apply.
iv. Under Axis, select Y.
v. Under Options, select Auto Range.
This will deactivate the Minimum and Maximum range fields.
vi. Click Apply, and Close the Axes - Force Monitor Plot panel.
(f) Close the Force Monitors panel.
10-46
3. Continue the calculation by requesting 70 time steps.

This performs the time-marching iterations for the next passing period, starting
from the current solution data at time = 0.1 second.
Note: Requesting 70 time steps will march the solution through 0.007 seconds, or
roughly one passing period. With the autosave and command monitor features
active (as defined above), the case, data, and pressure contour plot files will be
saved every 0.0005 seconds. The lift history should appear as in Figure 10.12.
-7.90e+00
-8.00e+00
-8.10e+00
Cl -8.20e+00
-8.30e+00
-8.40e+00
-8.50e+00
0.1 0.101 0.102 0.103 0.104 0.105 0.106 0.107
Time

Figure 10.12: Lift History During the Final Passing Period
4. Examine the results at different time steps within a single passing period.
(a) Read in the case and data files of interest.
(b) Display contours of static pressure.
The display of pressure contours every five time steps will show the time-
varying pressure distribution and the motion of the rotor. Examples of two
pressure contour plots at t = 0.1030 and t = 1.070 seconds are shown in
Figures 10.13 and 10.14, respectively.
Extra: If you generated a solution animation during the calculation as mentioned
earlier in this tutorial, you could play it back inside FLUENT to see the pressure
contour animation over time. See Tutorial 4 for details.

1.01e+05
1.01e+05
1.01e+05
1.00e+05
1.00e+05
9.97e+04
9.93e+04
9.90e+04
9.86e+04
9.82e+04
9.79e+04
9.75e+04
9.72e+04
9.68e+04
9.65e+04
9.61e+04
9.58e+04
9.54e+04
9.51e+04
9.47e+04
9.43e+04

Figure 10.13: Pressure Contours at Time = 0.103 Seconds
1.01e+05
1.01e+05
1.01e+05
1.00e+05
9.98e+04
9.94e+04
9.91e+04
9.87e+04
9.83e+04
9.79e+04
9.75e+04
9.71e+04
9.68e+04
9.64e+04
9.60e+04
9.56e+04
9.52e+04
9.48e+04
9.45e+04
9.41e+04
9.37e+04

Figure 10.14: Pressure Contours at Time = 0.107 Seconds
10-48
Summary
In this tutorial, you have modeled the time-periodic flow involving rotor-stator interac-
tion. You have learned how to merge the separate rotor and stator meshes using tmerge,
and to create the grid-interface zones along the sliding interface. Similar procedures can
be used to tie together meshes for non-sliding mesh analysis.
You have used FLUENT’s time-dependent flow prediction capability, and you have learned
how to set solution parameters for implicit time-stepping. These time-dependent flow
prediction procedures can also be applied to other, non-sliding mesh, analysis. The
procedures in this tutorial, however, are applicable to time-periodic calculations, in which
the initial condition and initial phase of the transient calculation are treated without
concern for time accuracy.
You have also learned how to manage the file saving and graphical postprocessing for
time-dependent flows, using file autosaving and command monitors to automatically save
solution information as the transient calculation proceeds.

10-50
Tutorial 11. Using Dynamic Meshes
Introduction
This tutorial provides information for performing basic dynamic mesh calculations. In
addition to combining the basic mesh-motion schemes, this tutorial will introduce rigid-
body motion of a cell zone. This is useful for a multitude of realistic cases with moving
meshes.
• Use the dynamic mesh capability of FLUENT to solve a simple flow-driven rigid-
body motion problem.
• Set boundary conditions for internal flow.
• Use a compiled user-defined function (UDF) to specify flow-driven rigid-body. mo-

tion
• Calculate a solution using the segregated solver.
Prerequisites
you have completed Tutorial 1. Some steps in the setup and solution procedure will not
Problem Description
The problem to be considered is shown schematically in Figure 11.1. A 2D axisymmetric
valve geometry is used, consisting of a pressurized cavity on the left, driving the motion
of a poppet that toggles the flow to the circumferential pressure outlets. A spring force
is also acting on the poppet. In this case the transient closure of the valve is studied.

Using Dynamic Meshes
pressure outlets
mass
flow
inlet
moving
poppet
Setup and Solution

Preparation
1. Download dynamic_mesh.zip from the Fluent Inc. User Services Center or copy
Tutorial 1).
2. Unzip dynamic_mesh.zip.
valve.msh and valve.c can be found in the /dynamic mesh folder created after
unzipping the file.
A user-defined function will be used to define the rigid-body motion of the poppet
in the valve geometry. This function has already been written (valve.c). You will
only need to compile it within FLUENT.
11-2
Step 1: Grid
1. Read the grid file valve.msh.
2. Check the grid.

Grid −→Check
Note: You should always make sure that the cell minimum volume is not negative,
since FLUENT cannot begin a calculation if this is the case.
3. Scale the grid.

Grid −→Scale...
(a) Under Unit Conversion, select in from the drop-down list to complete the phrase
Grid Was Created In in (inches).
(c) Click Change Length Units to set inches as the working units for length, and
then close the panel.


Grid
Figure 11.2: Initial Grid for the Valve
11-4
Step 2: Units
1. For convenience, define new units for pressure and mass flow.
In the problem description, pressure, length, and mass flow are specified in psi, in,
and gpm, respectively. While the units for length were switched while scaling the
grid in the previous step, psi and gpm are not the default units for pressure and
mass flow.
(a) Select pressure under Quantities, and psi under Units.

(b) Select mass-flow under Quantities, and click New...
The Define Unit panel will appear.
i. Enter gpm under Unit.

ii. Enter 0.0536265 under Factor.
iii. Click OK.

Step 3: Models
1. Enable an axisymmetric time-dependent calculation.
(a) Under Space, select Axisymmetric.

(b) Select Unsteady under Time.
(c) Keep the default Unsteady Formulation option of 1st-Order Implicit.
Dynamic mesh simulations currently work only with first-order time ad-
! vancement.
(d) Click OK.
11-6

(a) Select k-epsilon as the Model, and retain the default setting of Standard under
k-epsilon Model.
(b) Click OK.

Step 4: Materials
You will create a new material called oil.
1. In the Name field, enter oil.
2. Specify 850 for the Density.
3. Specify 0.17 for the Viscosity.
4. Click Change/Create.
5. Click No when prompted to overwrite air.
11-8

Set the operating pressure to 0 psi.
For this problem, you will work with absolute pressures.


Dynamic mesh motion and all related parameters are specified using the items in the
Define/Dynamic Mesh submenu, not through the Boundary Conditions panel. You will set
these conditions in the next step.
1. Set the conditions for the mass flow inlet (inlet) as shown in the following figure.
2. Click OK.
11-10
3. Set the conditions for the exit boundary (outlet) as shown in the following figure.
4. Set the conditions for the fluid zone, def zone.
5. Select oil as the Material Name from the drop-down list, and keep the default Motion
Type as Stationary.
This replaces air with oil as a working media.
6. Repeat this procedure for the other two fluid zones, inlet zone and outlet zone.


In this step, you will create a non-conformal grid interface between the deforming walls
corresponding to the radial boundary of the pressure outlets, and the deforming wall cor-
responding to the radial boundary of the deforming fluid zone next to the poppet.
1. Select ext intf in the Interface Zone 1 list.

Note that when one interface zone is smaller than the other, it is recommended that
you choose the smaller zone as Interface Zone 1.
2. Select int int in the Interface Zone 2 list.
3. Enter the name if under Grid Interface.
4. Click Create.
Note: In the process of creating the grid interface, FLUENT creates two new wall
boundary zones: wall-22 and wall-23. You will not be able to display these
walls.
wall-22 is the non-overlapping region of the ext intf zone that results from
the intersection of the ext intf and int int boundary zones, and is listed un-
der Boundary Zone 1 in the Grid Interfaces panel.
wall-23 is the non-overlapping region of the int int zone that results from the
intersection of the two interface zones, and is listed under Boundary Zone 2 in
the Grid Interfaces panel.
11-12
Note that in general, you will need to set boundary conditions for these new wall
zones, when they are not empty. In this case, default settings are used.
Step 8: Mesh Motion

1. Select and compile the user-defined function (UDF).
Define −→ User-Defined −→ Functions −→Compiled...
(a) Under Source Files, click Add...

A Select File panel will open.
(b) In the Select File panel, select the source code valve.c, and click OK.
(c) In the Compiled UDFs panel, click Build.
The UDF has already been defined, but it needs to be compiled within FLUENT
before it can be used in the solver. Here you will create a library with the
default name of libudf in your working directory. If you would like to use a
different name, you can enter it in the Library Name field. Note that in this
case you need to make sure that you will open the correct library in the next
step.
(d) Click OK in the dialog box that will appear.
(e) Click Load to load the UDF library you just compiled.
Once the UDF is built and loaded, it is now available to hook to your model. Its
name will appear as valve::libudf and may be selected in drop-down lists of various
panels.
2. Hook the reader and writer udf.

Define −→ User-Defined −→Function Hooks...

(a) Select reader::libudf in the drop-down list next to Read Data.

(b) Select writer::libudf in the drop-down list next to Write Data.
These two functions will read/write the position of C.G. and velocity in the X
direction to the data file. The location of C.G. and the velocity are necessary
for restarting a case. When starting from an intermediate case and data file,
FLUENT needs to know the location of C.G. and velocity, which are the initial
conditions for the motion calculation. Those values are saved in the data file
using the writer UDF and will be read in using the reader UDF when reading
the data file.
(c) Click OK to close the panel.
11-14
3. Activate dynamic mesh motion and specify the associated parameters.

Define −→ Dynamic Mesh −→Parameters...
(a) Under Models, select Dynamic Mesh.

See Chapter 10 of the User’s Guide for more information the available models
for moving and deforming zones.
(b) Under Mesh Methods, select Smoothing and Remeshing.
FLUENT will automatically flag the existing mesh zones for use of the different
dynamic mesh methods where applicable.
(c) Set the parameters under Smoothing as follows:
i. Keep the default specification of 1 for the Spring Constant Factor.
ii. Specify 0.7 for the Boundary Node Relaxation.
iii. Keep the default specification of 0.001 for the Convergence Tolerance.
iv. Specify 50 for the Number of Iterations.

(d) Set the parameters under Remeshing as follows:
i. Under Options, be sure that the Must Improve Skewness option is selected.
ii. Specify other parameters as shown the following table:
Parameter Value
Minimum Length Scale 2.0e-5
Maximum Length Scale 0.007
Maximum Cell Skewness 0.7
Size Remesh Interval 1
If a cell exceeds these limits, the cell is marked for remeshing. Therefore,
you will always need to specify problem-specific values under Remeshing
Parameters.
(e) Click OK.
4. Specify the motion of the poppet, the adjacent walls, and the fluid region left of
the poppet.
The poppet motion and the motion of the deforming wall side-wall-3 are specified by
means of the UDF valve.
Define −→ Dynamic Mesh −→Zones...
11-16
(a) Specify the motion of the poppet.
i. In the Zone Names drop-down list, select poppet.

ii. Under Type, keep the default selection of Rigid Body.
iii. Under Motion Attributes, select valve::libudf in the Motion UDF/Profile
drop-down list to hook the UDF to your model.
iv. Keep the default values of (0, 0) m for Center of Gravity Location, and 0
for Center of Gravity Orientation.
FLUENT will automatically update the position of the CG based on the
input you gave for the motion.
v. Click the Meshing Options tab.
vi. Specify a value of 0.005 for Cell Height.
vii. Click Create.

(b) Specify the motion of the deforming axis (def axis).
i. In the Zone Names drop-down list, select def axis.

ii. Under Type, select Deforming.
The panel will expand to show the inputs for a deforming zone.
iii. Click the Geometry Definition tab.
iv. In the Definition drop-down list, select plane.
The panel will expand again to show the inputs for a planar geometry.
v. Under Point on Plane, enter 0, 0.
vi. Under Plane Normal, enter 0, 1.
vii. Click the Meshing Options tab.
viii. Under Methods, keep the default selections of Smoothing and Remeshing.
ix. Specify a value of 0.0045 for Minimum Length Scale and a value of 0.0055
for Maximum Length Scale.
x. Keep the default value of 1 for Maximum Skewness.
xi. Click Create.
11-18
(c) Specify the motion of the deforming wall corresponding to the radial boundary
of the deforming fluid zone next to the poppet (int int).
i. In the Zone Names drop-down list, select int int.
ii. Under Type, keep the previous selection of Deforming.
iii. Click the Geometry Definition tab.
iv. In the Definition drop-down list, select plane.
The panel will expand again to show the inputs for a planar geometry.
v. Under Point on Plane, enter 0, 0.22625.
vi. Under Plane Normal, keep the previous setting of 0, 1.
vii. Click the Meshing Options tab.
viii. Under Methods, be sure that Smoothing and Remeshing are selected, and
keep the previous settings for Maximum Skewness to value of 1.
ix. Specify a value of 0.0045 for Minimum Length Scale and a value of 0.0055
for Maximum Length Scale.
x. Click Create.
In many MDM problems, you may want to preview the mesh motion before proceed-
ing any further. In this problem, the mesh motion is driven by the pressure exerted
by the fluid on the poppet and acting against the inertia of the poppet and the force
of a preloaded spring attached to it. Hence, for this problem, mesh motion in the
absence of a flow field solution is meaningless, and you will not use this feature
here.

Step 9: Solution
(a) Keep all default discretization schemes and values for under-relaxation factors.
This problem has been found to converge satisfactorily with these default set-
tings. Alternatively, you may want to try the PISO discretization scheme for
Pressure-Velocity Coupling in conjunction with higher under-relaxation factors
in order to reduce the overall CPU time needed for this simulation.
(b) Click OK.

3. Request that case and data files are automatically saved every 10 time steps.
11-20
To retain all files, keep the Overwrite Existing Files inactive.
(b) In the Filename field, enter valve.
When FLUENT saves a file, it will append the time step value to the file name
prefix (valve). The standard extensions (.cas and .dat) will also be ap-
pended.
(c) Click OK.

You will initialize the flow field at this point in order to be able to display contours
and vectors that you will use to define animations.

(a) Set Gauge Pressure to 80 psi.

(b) Set Axial Velocity to 3.097237 m/s.
(c) Set Turbulence Kinetic Energy to 0.1438932.
(d) Set Turbulence Dissipation Rate to 16.8147.
(e) Click Init.
5. Create animation sequences for the static pressure contour plots and velocity vectors
plots in the valve.
You will use FLUENT’s solution animation feature to save contour plots of temper-
ature every 5 time steps. After the calculation is complete, you will use the solution
animation playback feature to view the animated temperature plots over time.
Solve −→ Animate −→Define...
(a) Increase the number of Animation Sequences to 2.

(b) Under Name, enter pressure for the first animation, and vv for the second
one.
(c) Under Every, increase the number to 5 for both sequences.
The default value of 1 instructs FLUENT to update the animation sequence at
every time step. For this case, this would generate a large number of files.
(d) In the When drop-down list, select Time Step.
11-22
(e) Define the animation sequence for pressure.

i. Click Define... on the line for pressure to set the parameters for the
sequence.
ii. Under Storage Type, keep the default selection of Metafile.

Note: If you want to store the plots in a directory other than your working
directory, enter the directory path in the Storage Directory field. If
this field is blank (the default), the files will be saved in your working
directory (i.e., the directory where you started FLUENT).
iii. Increase the Window number to 1 and click Set.
iv. Under Display Type, select Contours.
The Contours panel will open.

A. Under Options, turn on Filled.

B. In the Contours of drop-down lists, select Pressure... and Static Pres-
sure.
C. Click Display.
8.00e+01
7.67e+01
7.35e+01
7.02e+01
6.69e+01
6.37e+01
6.04e+01
5.71e+01
5.39e+01
5.06e+01
4.74e+01
4.41e+01
4.08e+01
3.76e+01
3.43e+01
3.10e+01
2.78e+01
2.45e+01
2.12e+01
1.80e+01
1.47e+01
Contours of Static Pressure (psi) (Time=0.0000e+00)

FLUENT 6.2 (axi, segregated, dynamesh, ske, unsteady)
Figure 11.3: Contours of Static Pressure at t = 0 s
11-24
v. Click OK in the Animation Sequence panel.

The Animation Sequence panel will close, and the checkbox in the Active
column next to pressure in the Solution Animation panel will become se-
lected.
(f) Define the animation sequence for the velocity vectors.
i. Click Define... on the line for vv to set the parameters for the sequence.
ii. Under Storage Type, keep the default selection of Metafile.
iii. Increase the Window number to 2 and click Set.
iv. Under Display Type, select Vectors.
The Vectors panel will open.
v. Click Display in the Vectors panel.

vi. Click OK in the Animation Sequence panel.
The Animation Sequence panel will close, and the checkbox in the Active
column next to vv in the Solution Animation panel will become selected.

3.10e+00
3.10e+00
3.10e+00
3.10e+00
3.10e+00
3.10e+00
3.10e+00
3.10e+00
3.10e+00
3.10e+00
3.10e+00
3.10e+00
3.10e+00
3.10e+00
3.10e+00
3.10e+00
3.10e+00
3.10e+00
3.10e+00
3.10e+00
3.10e+00
Velocity Vectors Colored By Velocity Magnitude (m/s) (Time=0.0000e+00)

Figure 11.4: Vectors of Velocity at t = 0 s
vii. Click OK in the Solution Animation panel.
6. Set the time step parameters for the calculation.

(a) Set the Time Step Size to 2e-6 s.
(b) Increase the Max Iterations per Time Step to 40.
In the accurate solution of a real-life time-dependent CFD problem, it is impor-
tant to make sure that the solution converges at every time step to within the
desired accuracy. Here the first few time steps will only come to a reasonably
converged solution.
(c) Click Apply.
This will save the time step size to the case file (the next time a case file is
saved).
11-26
7. Save the initial case and data files (valve.cas and valve.dat).
8. Request 80 time steps.

Extra: If you decide to read in the case file that is provided for this tutorial on the
documentation CD, you will need to compile the UDF associated with this tutorial
in your working directory. This is necessary because FLUENT will expect to find
the correct UDF libraries in your working directory when reading the case file.
The UDF (valve.c) that is provided can be edited and customized by changing the
parameters as required for your case. In this tutorial, the values necessary for this
case were preset in the source code. These values may be modified to best suit your
model.


1. Inspect the solution at the final time step.
(a) Inspect the contours of static pressure in the valve (Figure 11.5).
The negative absolute pressure indicates cavitating flow.
See Chapter 24.6.4 of the User’s Guide for details about the cavitation model.
8.17e+02
7.60e+02
7.04e+02
6.47e+02
5.90e+02
5.33e+02
4.77e+02
4.20e+02
3.63e+02
3.07e+02
2.50e+02
1.93e+02
1.36e+02
7.97e+01
2.29e+01
-3.38e+01
-9.05e+01
-1.47e+02
-2.04e+02
-2.61e+02
-3.17e+02
Contours of Static Pressure (psi) (Time=1.6000e-04)

Figure 11.5: Contours of Static Pressure After 80 Time Steps
(b) Inspect the velocity vectors in the valve (Figure 11.6).
2. Optionally, inspect the solution at different intermediate time steps.

(a) Read in the corresponding case and data files (valve.cas and valve.dat).
(b) Display the desired contours and vectors.
11-28
7.77e+01
7.38e+01
6.99e+01
6.60e+01
6.21e+01
5.83e+01
5.44e+01
5.05e+01
4.66e+01
4.28e+01
3.89e+01
3.50e+01
3.11e+01
2.73e+01
2.34e+01
1.95e+01
1.56e+01
1.18e+01
7.88e+00
4.00e+00
1.23e-01

Figure 11.6: Velocity Vectors After 80 Time Steps
3. Play back the animation of the pressure contours.

(a) In the Sequences list, select pressure.

The playback control buttons will become active.

(b) Set the slider bar above Replay Speed about halfway in between Slow and Fast.
(c) Keep the default settings in the rest of the panel and click the play button
(the second from the right in the group of buttons under Playback).
See Tutorial 4 and Section 26.20 of the User’s Guide for additional information
on animating the solution.
4. Play back the animation of the velocity vectors.

(a) In the Sequences list, select vv.
(b) Keep the default settings in the rest of the panel and click the play button.
Summary
In this tutorial you learned how to use the dynamic mesh feature of FLUENT to simulate
the rigid-body motion of a valve poppet in a flow field, driven by the flow-generated
forces, and spring and inertial forces, by means of a user defined function (UDF).
11-30
Tutorial 12. Modeling Species Transport and Gaseous
Combustion
Introduction
This tutorial examines chemical species mixing and combustion of a gaseous fuel. A cylin-
drical combustor burning methane (CH4 ) in air is studied using the finite-rate chemistry
model in FLUENT.
• Enable physical models, select material properties, and define boundary conditions
for a turbulent flow with chemical species mixing and reaction
• Initiate and solve the combustion simulation using the segregated solver
• Compare the results computed with constant and variable specific heat
• Examine the reacting flow results using graphics
• Predict thermal and prompt NOx production
• Use custom field functions to compute NO parts per million
Prerequisites
You may find it helpful to read Chapter 13 of the User’s Guide to learn more about
chemical reaction modeling. Otherwise, no previous experience with chemical reaction
or combustion modeling is assumed.
Problem Description
The cylindrical combustor considered in this tutorial is shown in Figure 12.1. The flame
considered is a turbulent diffusion flame. A small nozzle in the center of the combustor
introduces methane at 80 m/s. Ambient air enters the combustor coaxially at 0.5 m/s.

Modeling Species Transport and Gaseous Combustion
The overall equivalence ratio is approximately 0.76 (about 28% excess air). The high-
speed methane jet initially expands with little interference from the outer wall, and
entrains and mixes with the low-speed air. The Reynolds number based on the methane
jet diameter is approximately 5.7 × 103 .
Air, 0.5 m/s, 300K Wall: 300K
0.225 m
0.005m
Methane, 80 m/s, 300K

L
C
1.8 m
Figure 12.1: Combustion of Methane Gas in a Turbulent Diffusion Flame Furnace
12-2
Background
In this tutorial, you will use the generalized finite-rate chemistry model to analyze the
methane-air combustion system. The combustion will be modeled using a global one-
step reaction mechanism, assuming complete conversion of the fuel to CO2 and H2 O. The
reaction equation is
CH4 + 2O2 → CO2 + 2H2 O (12.1)
This reaction will be defined in terms of stoichiometric coefficients, formation enthalpies,

and parameters that control the reaction rate. The reaction rate will be determined as-
suming that turbulent mixing is the rate-limiting process, with the turbulence-chemistry
interaction modeled using the eddy-dissipation model.
Setup and Solution

Preparation
1. Download species_transport.zip from the Fluent Inc. User Services Center or
2. Unzip species_transport.zip.
gascomb.msh can be found in the /species transport folder created after unzip-
ping the file.

Step 1: Grid
1. Read the grid file gascomb.msh.
After reading the grid file, FLUENT will report that 1615 quadrilateral fluid cells
have been read, along with a number of boundary faces with different zone identi-
fiers.
2. Check the grid.

Grid −→Check
The grid check lists the minimum and maximum x and y values from the grid,
and reports on a number of other grid features that are checked. Any errors in the
grid would be reported at this time. For instance, the cell volumes must never be
negative. Note that the domain extents are reported in units of meters, the default
unit of length in FLUENT. Since this grid was created in units of millimeters, the
Scale Grid panel will be used to scale the grid into meters.
3. Scale the grid.

Grid −→Scale...
(a) Under Unit Conversion, select mm from the drop-down list to complete the
phrase Grid Was Created In mm.
(b) Click on Scale and confirm that the maximum x and y values are 1.8 and 0.225
meters, respectively, as indicated in Figure 12.1.
Note: Because the default SI units will be used in this tutorial, there is no need to
change any units in this problem.

Extra: You can use the right mouse button to check which zone number corresponds
to each boundary. If you click the right mouse button on one of the boundaries
in the graphics window, its zone number, name, and type will be printed in the
FLUENTconsole window. This feature is especially useful when you have several
zones of the same type and you want to distinguish between them quickly.
12-4
Grid
Figure 12.2: The Quadrilateral Grid for the Combustor Model

Step 2: Models
1. Define the domain as axisymmetric, and keep the default (segregated) solver.
2. Enable the k- turbulence model.

12-6
The panel will expand to provide further options. Click OK to accept the default
Standard model and parameters.


4. Enable chemical species transport and reaction.

Define −→ Models −→ Species −→Transport & Reaction...
(a) Select Species Transport under Model.

(b) Select Volumetric under Reactions.
(c) Choose methane-air in the Mixture Material drop-down list.
The Mixture Material list contains the set of chemical mixtures that exist in
the FLUENT database. By selecting one of the pre-defined mixtures, you are
accessing a complete description of the reacting system. The chemical species
in the system and their physical and thermodynamic properties are defined by
your selection of the mixture material. You can alter the mixture material
selection or modify the mixture material properties using the Materials panel
(see Step 3: Materials).
12-8
(d) Select Eddy-Dissipation under Turbulence-Chemistry Interaction.

The eddy-dissipation model computes the rate of reaction under the assumption
that chemical kinetics are fast compared to the rate at which reactants are
mixed by turbulent fluctuations (eddies).
(e) Click OK.
After you click OK in the Species Model panel, a warning about the symmetry
zone will appear in the console window:
Warning: It appears that symmetry zone 5 should actually be an axis

(it has faces with zero area projections).
Unless you change the zone type from symmetry to axis,
you may not be able to continue the solution without
encountering floating point errors.
In this axisymmetric model, the centerline should be treated using the axis
boundary condition instead of symmetry. You will change the symmetry zone
to an axis boundary in Step 4: Boundary Conditions.
The console window will also list the properties that are required for the models
you have enabled. You will see an Information dialog box, reminding you to
confirm the property values that have been extracted from the database.
(f) Click OK to continue.

Step 3: Materials
The Materials panel shows the mixture material, methane-air, that was enabled in the
Species Model panel. The properties for this mixture material have been copied from the
FLUENT database and can be modified by you.
Here, you will modify the default setting for the mixture by enabling the gas law. By
default, the mixture material uses constant properties: you will retain this constant prop-
erty assumption for now, allowing only the mixture density to vary with temperature and
composition. The influence of variable property inputs on the combustion prediction will
be examined in a later part of this tutorial.
1. Retain incompressible-ideal-gas in the Density drop-down list.
2. Click the Edit... button to the right of Mixture Species.

This opens the Species panel.
12-10
You can add or remove species from the mixture material using this panel. Here,
the species that make up the methane-air mixture are predefined and require no
modification.
3. Click Cancel to close the panel without making any changes.
4. In the Materials panel, click the Edit... button to the right of the Reaction drop-
down list.
This will open the Reactions panel.

The eddy-dissipation reaction model ignores chemical kinetics (the Arrhenius rate)
and uses only the Mixing Rate parameters in the Reactions panel. The Arrhenius
Rate section of the panel is therefore inactive. (The Rate Exponent and Arrhenius
Rate entries are included in the database and are employed when the alternate finite-
rate/eddy-dissipation model is used.) See the User’s Guide for details.
5. Accept the default settings for the Mixing Rate constants by clicking the OK button.
6. In the Materials panel, select constant from the drop-down list next to Cp and enter
1000 for the specific heat value.
7. Use the scroll bar to review the remaining properties. Click on the Change/Create
button to accept the material property settings and then Close the panel.
12-12
As noted above, the initial calculation will be performed assuming that all properties except
density are constant. Using constant transport properties (viscosity, thermal conductivity,
and mass diffusion coefficients) is acceptable here because the flow is fully turbulent. The
molecular transport properties will play a minor role compared to turbulent transport. The
assumption of constant specific heat, in contrast, has a strong effect on the combustion
solution, and you will change this property definition in Step 6: Solution Using Non-
Constant Heat Capacity.


1. Convert the symmetry zone to the axis type.
The symmetry zone must be converted to an axis to prevent numerical difficulties
where the radius goes to zero.
(a) Select symmetry-5 in the Zone list and then select axis in the Type list.
You will be prompted to accept the change of boundary type:
(b) Click Yes to confirm the change.
(c) In the resulting Axis panel, click OK to accept the default axis zone name.
12-14
2. Set the boundary conditions for the air inlet, velocity-inlet-8.

Hint: Redisplay the grid without the fluid zone. This will show the boundaries.
Use the right mouse button to probe the air inlet. The console window and the
Boundary Conditions panel will show that the air inlet is labeled velocity-inlet-8.
(a) Rename the boundary as air-inlet in the Zone Name text entry box.
(b) Set the boundary conditions at the air inlet as shown in the panel.

3. Set the fuel inlet boundary conditions for velocity-inlet-6.
(a) Rename this zone fuel-inlet and assign inlet conditions as shown in the
panel.
12-16
4. Set the following conditions for the exit boundary, pressure-outlet-9:
Note: The Backflow values in this panel are utilized only when backflow occurs
at the pressure outlet. Reasonable values should always be assigned, since
backflow may occur during intermediate iterations and could affect the solution
stability.

5. Set the boundary conditions for the outer wall, wall-7.

Hint: Use the mouse-probe method described above for the air inlet to determine
which zone corresponds to the outer wall. The outer wall zone will be selected
in the Boundary Conditions panel once the outer wall boundary is probed.
(a) Rename this boundary outer-wall in the Zone Name text entry box.
(b) Set the thermal condition to Temperature and keep the default temperature of
300 K.
(c) Retain the default settings in the Momentum and Species sections of the panel.
12-18
6. Set the boundary conditions for wall-2, which represents the small fuel inlet nozzle.
(a) Rename this boundary nozzle in the Zone Name text entry box.
(b) Accept the default thermal condition of Heat Flux with a value of zero (adia-
batic wall).
(c) Retain the default settings in the Momentum and Species sections of the panel.

Step 5: Initial Solution Using Constant Heat Capacity

1. Initialize the field variables.
(a) Select all-zones in the Compute From drop-down list.

(b) Adjust the Initial Values for Temperature to 2000 and ch4 Mass Fraction to 0.2.
(c) Click Init to initialize the variables, and then close the panel.
Initializing the flow using a high temperature and non-zero fuel content will
allow the combustion reaction to begin. The initial condition acts as a numer-
ical “spark” to ignite the methane-air mixture. This initialization is especially
critical when you include finite-rate kinetics in the overall reaction rate.
2. Set the under-relaxation factors.

The default under-relaxation parameters in FLUENT are set to high values. For
a combustion model it may be necessary to reduce the under-relaxation to stabilize
the solution. Some experimentation is typically necessary to establish the opti-
mal under-relaxation. For this tutorial it is sufficient to reduce the species under-
relaxation to 0.9.
12-20
(a) Use the slider bar next to the Under-Relaxation Factors list to locate each
species and set the under-relaxation factor for all the species to 0.9.



(b) Click OK.
4. Save the case file (gascomb1.cas).
(a) Keep the Write Binary Files button on to produce a smaller, unformatted binary
file.
(b) Enter the file name gascomb1.cas in the Case File text entry box.
(c) Click OK to proceed with the file writing.
12-22
The solution converges in about 300 iterations.
6. Save the case and data files (gascomb1.cas and gascomb1.dat).

Note: FLUENT will ask you to confirm that the previous case file is to be overwrit-
ten.
7. Review the current state of the solution by viewing contours of temperature (Fig-
ure 12.3).

list.
(b) Click Display.
The temperature contours are shown in Figure 12.3. The peak temperature, pre-
dicted using a constant heat capacity of 1000 J/kg-K, is over 3000 K. This over-
prediction of the flame temperature can be remedied by a more realistic model for
the temperature and composition dependence of the heat capacity, as illustrated in
the next step of the tutorial.

3.06e+03
2.92e+03
2.78e+03
2.65e+03
2.51e+03
2.37e+03
2.23e+03
2.09e+03
1.96e+03
1.82e+03
1.68e+03
1.54e+03
1.40e+03
1.27e+03
1.13e+03
9.90e+02
8.52e+02
7.14e+02
5.76e+02
4.38e+02
3.00e+02

FLUENT 6.2 (axi, segregated, spe, ske)
Figure 12.3: Temperature Contours: Constant cp
12-24
Step 6: Solution Using Non-Constant Heat Capacity

As noted above, the strong temperature and composition dependence of the specific heat
will have a significant impact on the predicted flame temperature. In this step you will
use the temperature-varying property information in the FLUENT database to recompute
the solution.
1. Enable composition dependence of the specific heat.

(a) In the drop-down list next to Cp, select mixing-law as the specific heat method.
(b) Click on the Change/Create button to render the mixture specific heat based
on a local mass-fraction-weighted average of all the species.
2. Enable temperature dependence of the specific heat for each species.

(a) In the Material Type drop-down list, select fluid.

The fluid material type gives you access to each species in the mixture.
(b) Select carbon-dioxide (co2) under Fluent Fluid Materials.
(c) In the drop-down list for Cp, select piecewise-polynomial.
This will open the Piecewise-Polynomial Profile panel.
12-26
i. Click OK to accept the default coefficients describing the temperature

variation of cp for carbon dioxide.
The default coefficients describe the polynomial cp (T ) and are extracted
from the FLUENT property database.
(d) Click on Change/Create in the Materials panel to accept the change in proper-
ties for CO2 .
(e) Repeat steps (b) and (c) above for the remaining species (CH4 , N2 , O2 , and
H2 O). Remember to click on Change/Create to accept the change for each
species.

Note: The residuals will jump significantly as the solution adjusts to the new spe-
cific heat representation. The solution converges after about 230 additional
iterations.
4. Save the new case and data files (gascomb2.cas and gascomb2.dat).

Review the solution by examining graphical displays of the results and performing surface
integrations at the combustor exit.
1. View contours of temperature (Figure 12.4).

list.
(b) Click Display.
The temperature contours are shown in Figure 12.4. The peak temperature has
dropped to about 2300 K as a result of the temperature- and composition-dependent
specific heat.
2.30e+03
2.20e+03
2.10e+03
2.00e+03
1.90e+03
1.80e+03
1.70e+03
1.60e+03
1.50e+03
1.40e+03
1.30e+03
1.20e+03
1.10e+03
9.99e+02
8.99e+02
8.00e+02
7.00e+02
6.00e+02
5.00e+02
4.00e+02
3.00e+02

Figure 12.4: Temperature Contours: Variable cp
2. Plot contours of specific heat (Figure 12.5).

Contours of the mixture specific heat will show how it varies through the domain.
(a) Select Properties... and Specific Heat (Cp) in the Contours of drop-down list.
(b) Click Display.
12-28
The contours are shown in Figure 12.5. The mixture specific heat is largest where
the CH4 is concentrated, near the fuel inlet, and where the temperature and com-
bustion product concentrations are large. The increase in heat capacity, relative to
the constant value used before, substantially lowers the peak flame temperature.
2.77e+03
2.68e+03
2.59e+03
2.50e+03
2.41e+03
2.33e+03
2.24e+03
2.15e+03
2.06e+03
1.98e+03
1.89e+03
1.80e+03
1.71e+03
1.62e+03
1.54e+03
1.45e+03
1.36e+03
1.27e+03
1.19e+03
1.10e+03
1.01e+03
Contours of Specific Heat (Cp) (j/kg-k)

Figure 12.5: Contours of Specific Heat


(a) Click the Vector Options... button.

This opens the Vector Options panel.
(b) Select the Fixed Length option and click Apply.

The fixed length option is useful when the vector magnitude varies dramatically.
With fixed length vectors, the velocity magnitude is described only by color
instead of by both vector length and color.
12-30
(c) In the Vectors panel, reset the Scale to 0.01 and click Display.
The velocity vectors are shown in Figure 12.6.
9.26e+01
8.80e+01
8.34e+01
7.88e+01
7.42e+01
6.96e+01
6.50e+01
6.03e+01
5.57e+01
5.11e+01
4.65e+01
4.19e+01
3.73e+01
3.26e+01
2.80e+01
2.34e+01
1.88e+01
1.42e+01
9.57e+00
4.95e+00
3.38e-01

Figure 12.6: Velocity Vectors: Variable cp
4. Plot contours of stream function (Figure 12.7).


(a) Select Velocity... and Stream Function in the Contours of drop-down list.
(b) Click Display.
The stream function contours are shown in Figure 12.7. The entrainment of air
into the high-velocity methane jet is clearly visible in the streamline display.
5. Plot contours of mass fraction for each species.

(a) Select Species... and Mass fraction of ch4 in the Contours of drop-down list.
(b) Turn on the Filled button under Options.
(c) Click Display.
The CH4 mass fraction contours are shown in Figure 12.8.
(d) Repeat for the remaining species.
The mass fraction contours for O2 , CO2 , and H2 O are shown in Figures 12.9, 12.10,
and 12.11.
12-32
1.55e-02
1.48e-02
1.40e-02
1.32e-02
1.24e-02
1.17e-02
1.09e-02
1.01e-02
9.32e-03
8.55e-03
7.77e-03
6.99e-03
6.21e-03
5.44e-03
4.66e-03
3.88e-03
3.11e-03
2.33e-03
1.55e-03
7.77e-04
0.00e+00

Figure 12.7: Stream Function Contours: Variable cp
1.00e+00
9.50e-01
9.00e-01
8.50e-01
8.00e-01
7.50e-01
7.00e-01
6.50e-01
6.00e-01
5.50e-01
5.00e-01
4.50e-01
4.00e-01
3.50e-01
3.00e-01
2.50e-01
2.00e-01
1.50e-01
1.00e-01
5.00e-02
0.00e+00
Contours of Mass fraction of ch4

Figure 12.8: CH4 Mass Fraction

2.30e-01
2.19e-01
2.07e-01
1.96e-01
1.84e-01
1.73e-01
1.61e-01
1.50e-01
1.38e-01
1.27e-01
1.15e-01
1.04e-01
9.20e-02
8.05e-02
6.90e-02
5.75e-02
4.60e-02
3.45e-02
2.30e-02
1.15e-02
0.00e+00
Contours of Mass fraction of o2

Figure 12.9: O2 Mass Fraction
1.47e-01
1.39e-01
1.32e-01
1.25e-01
1.17e-01
1.10e-01
1.03e-01
9.53e-02
8.80e-02
8.06e-02
7.33e-02
6.60e-02
5.86e-02
5.13e-02
4.40e-02
3.66e-02
2.93e-02
2.20e-02
1.47e-02
7.33e-03
0.00e+00
Contours of Mass fraction of co2

Figure 12.10: CO2 Mass Fraction
12-34
1.20e-01
1.14e-01
1.08e-01
1.02e-01
9.60e-02
9.00e-02
8.40e-02
7.80e-02
7.20e-02
6.60e-02
6.00e-02
5.40e-02
4.80e-02
4.20e-02
3.60e-02
3.00e-02
2.40e-02
1.80e-02
1.20e-02
6.00e-03
0.00e+00
Contours of Mass fraction of h2o

Figure 12.11: H2 O Mass Fraction

6. Determine the average exit temperature and velocity.

(a) Select Mass-Weighted Average in the Report Type drop-down list.

(b) Select Temperature... and Static Temperature in the Field Variable drop-down
list.
The mass-averaged temperature will be computed as
T ρ~v · dA
R
~
T = R (12.2)
ρ~v · dA~
(c) Select pressure-outlet-9 as the surface over which to perform the integration.
(d) Click Compute.
The mass-weighted average exit temperature is about 1796 K.
12-36
(e) Select Area-Weighted Average as the Report Type and Velocity Magnitude as
the Field Variable.
The area-weighted velocity-magnitude average will be computed as
1Z
v̄ = v dA (12.3)
A
(f) Click Compute.
The area-averaged exit velocity is about 3.14 m/s.

Step 8: NOx Prediction

In this section you will extend the FLUENT model to include the prediction of NOx . You
will first calculate the formation of both thermal and prompt NOx , then calculate each
separately to determine the contribution of each mechanism.
1. Enable the NOx model.

Define −→ Models −→ Species −→NOx...
(a) Under Pathways, enable Thermal NO and Prompt NO.

(b) Click the Turbulence Interaction tab and select Temperature in the PDF Mode
drop-down list under Turbulence Interaction Mode.
This will enable the turbulence-chemistry interaction.
If turbulence interaction is not enabled, you will be computing NOx formation
without considering the important influence of turbulent fluctuations on the
time-averaged reaction rates.
(c) Click the Thermal tab under Formation Model Parameters.
(d) Select partial-equilibrium in the [O] Model drop down list.
The partial-equilibrium model is used to predict the O radical concentration
required for thermal NOx prediction.
(e) Click the Prompt tab under Formation Model Parameters.
(f) Set the Equivalence Ratio to 0.76, and keep the default Fuel Species and Fuel
Carbon Number.
The equivalence ratio defines the fuel-air ratio (relative to stoichiometric con-
ditions) and is used in the calculation of prompt NOx formation. The Fuel
12-38
Carbon Number is the number of carbon atoms per molecule of fuel and is used
in the prompt NOx prediction. The Fuel Species designation is also used in the
prompt NOx model.
(g) Click OK to accept these changes.
2. Enable the calculation of only the NO species, and set the under-relaxation factor
for this equation.
(a) In the Equations list, deselect all variables except the NO species.
(b) Increase the NO under-relaxation factor to 1.0.
You will predict NOx formation in a “postprocessing” mode, with the flow field,
temperature, and hydrocarbon combustion species concentrations fixed. Thus, only
the NO equation is computed. Prediction of NO in this mode is justified on the
grounds that the NO concentrations are very low and have negligible impact on the
hydrocarbon combustion prediction.
3. Reduce the convergence criterion for the NO species equation.


(a) Set the Convergence Criterion to 1e-6 and click OK.

5. Save the new case and data files (gascomb3.cas and gascomb3.dat).
6. Review the solution by displaying contours of NO mass fraction (Figure 12.12).

(a) Select NOx... and Mass fraction of NO in the Contours of drop-down list.
(b) Deselect Filled under Options and click Display.
The NO mass fraction contours are shown in Figure 12.12. The peak con-
centration of NO is located in a region of high temperature where oxygen and
nitrogen are available.
7. Calculate the average exit NO mass fraction.

12-40
4.11e-03
3.90e-03
3.70e-03
3.49e-03
3.28e-03
3.08e-03
2.87e-03
2.67e-03
2.46e-03
2.26e-03
2.05e-03
1.85e-03
1.64e-03
1.44e-03
1.23e-03
1.03e-03
8.21e-04
6.16e-04
4.11e-04
2.05e-04
0.00e+00
Contours of Mass fraction of NO

Figure 12.12: Contours of NO Mass Fraction: Prompt and Thermal NOx
(a) Select Mass-Weighted Average in the Report Type drop-down list and NOx...
and Mass fraction of NO in the Field Variable drop-down list.
(b) Select pressure-outlet-9 as the surface over which to perform the integration.
(c) Click Compute.
The mass-weighted average exit NO mass fraction is about 0.00469.

8. Disable the prompt NOx mechanism and solve for thermal NOx only.
(a) Turn off Prompt NO under Models to disable the prompt NOx mechanism, and
click Apply.
(b) Request 50 iterations.
(c) Review the thermal NOx solution by viewing contours of NO mass fraction
(Figure 12.13).
i. Check that NOx... and Mass fraction of NO are selected in the Contours
of drop-down list.
ii. Click Display.
The NO mass fraction contours are shown in Figure 12.13. The concen-
tration of NO is slightly lower without the prompt NOx mechanism.
4.07e-03
3.87e-03
3.67e-03
3.46e-03
3.26e-03
3.06e-03
2.85e-03
2.65e-03
2.44e-03
2.24e-03
2.04e-03
1.83e-03
1.63e-03
1.43e-03
1.22e-03
1.02e-03
8.15e-04
6.11e-04
4.07e-04
2.04e-04
3.19e-20

Figure 12.13: Contours of NO Mass Fraction: Thermal NOx Formation
(d) Compute the average exit NO mass fraction with only thermal NOx formation.

Hint: Follow the same procedure you used earlier for the calculation with both
thermal and prompt NOx formation.
12-42
The mass-weighted average exit NO mass fraction, with thermal but no prompt
NOx formation, is about 0.00457.
9. Solve for prompt NOx production only.
(a) Turn off Thermal NO and turn on Prompt NO under Models, and click Apply.
(b) Request 50 iterations.
(c) Review the prompt NOx solution by viewing contours of NO mass fraction
(Figure 12.14).
The NO mass fraction contours are shown in Figure 12.14. The prompt NOx
mechanism is most significant in fuel-rich flames. In this case the flame is
lean and prompt NO production is low.
6.98e-05
6.63e-05
6.28e-05
5.93e-05
5.58e-05
5.23e-05
4.88e-05
4.53e-05
4.19e-05
3.84e-05
3.49e-05
3.14e-05
2.79e-05
2.44e-05
2.09e-05
1.74e-05
1.40e-05
1.05e-05
6.98e-06
3.49e-06
6.04e-27

Figure 12.14: Contours of NO Mass Fraction: Prompt NOx Formation
(d) Compute the average exit NO mass fraction with only prompt NOx formation.
Hint: Follow the same procedure you used earlier for the calculation with both
thermal and prompt NOx formation.
The mass-weighted average exit NO mass fraction, with only prompt NOx for-
mation, is about 0.000071.

Note: The individual thermal and prompt NO mass fractions do not add up
to the levels predicted with the two models combined. This is because
reversible reactions are involved. NO produced in one reaction can be
destroyed in another reaction.
10. Use a custom field function to compute NO parts per million (ppm).
Define −→Custom Field Functions...
NO ppm is computed from the following equation:
NO mole fraction × 106

NO ppm = (12.4)
1 − H2 O mole fraction
where
NO mass fraction × mixture MW

NO mole fraction = (12.5)
30
and the mixture molecular weight is
1
mixture MW = X mass fraction (12.6)
MW
i
where MW is the molecular weight of each species.

First you will create a function for Equation 12.6. Then you will substitute Equa-
tion 12.5 into Equation 12.4 and create a function for Equation 12.4.
(a) Create a custom field function for the mixture molecular weight.
12-44
i. Click on the 1 calculator button, then on /, and then on (.

ii. Select Species... and Mass fraction of ch4 in the Field Functions drop-down
list. Click Select to add this variable to the field function Definition.
iii. Click on / and then click on 1 and 6 to enter 16 (the molecular weight of
methane).
iv. Continue in this fashion to complete the definition of the mixture molec-
ular weight field function.
v. Enter bulk-mw in the New Function Name text entry box.
vi. Click Define to add the new field function to the variable list.
(b) Create a field function for NO ppm.
i. Select NOx... and Mass fraction of NO in the Field Functions drop-down

list. Click Select to add this variable to the field function Definition.
ii. Click the × button to introduce the multiplication sign.
iii. Select Custom Field Functions... and bulk-mw in the Field Functions drop-
down list. Click Select to add this variable to the field function Definition.
iv. Click on / and then click on 3 and 0 to enter 30 (the molecular weight of
NO).
v. Click the × button and then click on 1 and 0 to enter 10.
vi. Click on y^x and then on 6.
vii. Complete the definition of NO ppm as shown in the panel above.
viii. Enter no-ppm in the New Function Name text entry box.
ix. Click Define to add the new field function to the variable list.

11. Plot contours of NO ppm (Figure 12.15).

(a) Select Custom Field Functions... and no-ppm in the Contours of drop-down list.
(b) Click Display.
The NO ppm contours are shown in Figure 12.15. The contours closely re-
semble the mass fraction contours (Figure 12.14), as expected.
7.80e+01
7.41e+01
7.02e+01
6.63e+01
6.24e+01
5.85e+01
5.46e+01
5.07e+01
4.68e+01
4.29e+01
3.90e+01
3.51e+01
3.12e+01
2.73e+01
2.34e+01
1.95e+01
1.56e+01
1.17e+01
7.80e+00
3.90e+00
5.80e-21
Contours of no-ppm
Figure 12.15: Contours of NO ppm: Prompt NOx Formation
12-46
Summary
In this tutorial you used FLUENT to model the transport, mixing, and reaction of chemical
species. The reaction system was defined by using and modifying a mixture-material
entry in the FLUENT database. The procedures used here for simulation of hydrocarbon
combustion can be applied to other reacting flow systems.
This exercise illustrated the important role of the mixture heat capacity in the prediction
of flame temperature. The combustion modeling results are summarized in the following
table. (Note that some of the values in the table were not explicitly calculated during
the tutorial.)
Peak Temp. Exit Temp. Exit Velocity

(K) (K) (m/s)
Constant cp 3077 2198 3.83
Variable cp 2301 1796 3.14
The use of a constant cp results in a significant overprediction of the peak temperature.

The average exit temperature and velocity are also overpredicted.
While the variable cp solution produces dramatic improvements in the predicted results,
further improvements are possible by considering additional models and features available
in FLUENT, as discussed below.
The NOx production in this case was dominated by the thermal NO mechanism. This
mechanism is very sensitive to temperature. Every effort should be made to ensure that
the temperature solution is not overpredicted, since this will lead to unrealistically high
predicted levels of NO.
Further Improvements
Further improvements can be expected by including the effects of intermediate species
and radiation, both of which will result in lower predicted combustion temperatures.
The single-step reaction process used in this tutorial cannot account for the moderating
effects of intermediate reaction products, such as CO and H2 . Multiple-step reactions
can be used to address these species. If a multi-step Magnussen model is used, consid-
erably more computational effort is required to solve for the additional species. Where
applicable, the non-premixed combustion model can be used to account for intermediate
species at a reduced computational cost.
See Chapter 15 of the User’s Guide for more details on the non-premixed combustion
model.
Radiation heat transfer tends to make the temperature distribution more uniform, thereby
lowering the peak temperature. In addition, radiation heat transfer to the wall can be

very significant (especially here, with the wall temperature set at 300 K). The large
influence of radiation can be anticipated by computing the Boltzmann number for the
flow:
(ρUcp )inlet convection

Bo = 3
∼
σTAF radiation
where σ is the Boltzmann constant (5.729×10−8 W/m2 -K4 ) and TAF is the adiabatic
flame temperature. For a quick estimate, assume ρ = 1 kg/m3 , U = 0.5 m/s, and
cp = 1000 J/kg-K (the majority of the inflow is air). Assume TAF = 2000 K. The
resulting Boltzmann number is Bo = 1.09, which shows that radiation is just about as
important as convection for this problem.
See Section 12.3 of the User’s Guide and Tutorial 5 for details on radiation modeling.
12-48
Tutorial 13. Using the Non-Premixed Combustion Model
Introduction
A pulverized coal combustion simulation involves modeling a continuous gas phase flow
field and its interaction with a discrete phase of coal particles. The coal particles, traveling
through the gas, will devolatilize and undergo char combustion, creating a source of fuel
for reaction in the gas phase.
The reaction can be modeled using either the species transport model or the non-premixed
combustion model. In this tutorial you will model a simplified coal combustion furnace
using the non-premixed combustion model for the reaction chemistry.
• Define inputs for modeling non-premixed combustion chemistry
• Prepare a Probability Density Function (PDF) table in FLUENT for pulverized coal
• Define a discrete second phase of coal particles
• Solve a simulation involving reacting discrete phase coal particles
• Use the P-1 radiation model with particle/radiation interaction
The non-premixed combustion model uses a modeling approach that solves transport
equations for one or two conserved scalars, the mixture fractions. Multiple chemical
species, including radicals and intermediate species, may be included in the problem
definition. Their concentrations will be derived from the predicted mixture fraction
distribution.
Property data for the species are accessed through a chemical database and turbulence-
chemistry interaction is modeled using a β-function for the PDF. See Chapter 15 of the
User’s Guide for details on the non-premixed combustion modeling approach.
Prerequisites

Using the Non-Premixed Combustion Model
Problem Description
The coal combustion system considered in this tutorial is a simple 10 m by 1 m two-
dimensional duct depicted in Figure 13.1. As the model is symmetric, only half of the
domain width is modeled.
The inlet of the 2D duct is split into two streams. A high-speed stream near the center
of the duct enters at 50 m/s and spans 0.125 m. The other stream enters at 15 m/s and
spans 0.375 m. Both streams are air at 1500 K. Coal particles enter the furnace near the
center of the high-speed stream with a mass flow rate of 0.1 kg/s (total flow rate in the
furnace is 0.2 kg/s).
The duct wall has a constant temperature of 1200 K. The Reynolds number based on
the inlet dimension and the average inlet velocity is about 100,000. Thus, the flow is
turbulent.
Details of the coal composition and size distribution are included in Step 3: Non Adiabatic
PDF Table and Step 5: Materials.
T = 1200 K
w
Air: 15 m/s, 1500 K
0.5 m
Coal Injection: 0.1 kg/s
0.125 m Air: 50 m/s, 1500 K
Symmetry
Plane
10 m
Figure 13.1: 2D Furnace with Pulverized Coal Combustion
13-2
Setup and Solution

Preparation
1. Download non_premix_combustion.zip from the Fluent Inc. User Services Center
2. Unzip non_premix_combustion.zip.
coal.msh can be found in the /non premix combustion folder created after unzip-
ping the file.
The mesh file, coal.msh is a quadrilateral mesh describing the system geometry
shown in Figure 13.1.
Step 1: Grid
1. Read the 2D mesh file, coal.msh.
The FLUENT console window reports that the mesh contains 1357 quadrilateral
cells.
2. Check the grid.

Grid −→Check
The grid check should not report any errors or negative volumes.

Due to the grid resolution and the size of the domain, you may find it more useful
to display just the outline, or to zoom in on various portions of the grid display.
You can use the mouse zoom button (middle button, by default) to zoom in to
the display and the mouse probe button (right button, by default) to find out the
boundary zone labels. As annotated in Figure 13.2, the upstream boundary contains
two velocity inlets (for the low-speed and high-speed air streams), the downstream
boundary is a pressure outlet, the top boundary is a wall, and the bottom boundary
is a symmetry plane.

velocity-inlet-2 wall-7
velocity-inlet-8 symmetry
Grid
Figure 13.2: 2D Coal Furnace Mesh Display with Annotated Boundary Types
13-4
Step 2: Models
1. Accept the default segregated solver.
The non-premixed combustion model is available only with the segregated solver.
2. Turn on the Energy Equation.

Since heat transfer occurs in the system considered here, you will have to solve the
energy equation.
3. Turn on the standard k-epsilon turbulence model.

Reynolds number of the flow is approximately 105 . Hence the flow is turbulent and
the high-Re k- model is suitable.

4. Turn on the P1 radiation model.

Coal particles can radiate significantly, and the P-1 radiation model is appropriate
for combustors larger than 1m, so the optical thickness is higher than 1.
13-6
5. Turn on the Non-Premixed Combustion model.

(a) Under Model, select Non-Premixed Combustion.

The panel will expand to show the related inputs. You will use this panel to
create the PDF table.
When you use the non-premixed combustion model, you need to create a PDF table.
This table contains information on the thermo-chemistry and its interaction with
turbulence. FLUENT interpolates the PDF during the solution of the non-premixed
combustion model.

Step 3: Non Adiabatic PDF Table

1. In the Species Model panel, under PDF Options, turn on the Create Table option.
This will update the panel to display the inputs for the creating the PDF table.
The Inlet Diffusion option enables the mixture fraction to diffuse out of the domain
through inlets and outlets.
2. Define the chemistry models in the Chemistry tab.

For single mixture fraction cases, the assumed shape β-function is used since it
accurately represents the experimentally observed PDFs of mixture fraction.
(a) Retain the default options, Equilibrium and Non Adiabatic.

In most non-premixed combustion simulations, the Equilibrium chemistry model
is recommended. The Laminar Flamelets option can model local chemical non-
equilibrium due to turbulent strain, such as radical super-equilibrium, but can-
not model slow chemistry like NOx.
The coal combustor considered in this tutorial is a non-adiabatic system, with
heat transfer at the combustor wall and heat transfer to the coal particles from
the gas. Therefore, a non-adiabatic combustion system must be considered
while creating a PDF table.
(b) Retain the default value for Operating Pressure.
(c) Under Options, turn on Empirical Fuel Stream.
13-8
You can define either a single fuel stream, or a fuel stream plus a secondary
stream. Enabling a secondary stream allows you to keep track of two mixture
fractions. For coal combustion, this will allow you to track volatile matter (the
secondary stream) separately from the char (fuel stream). However, the two
mixture fraction has a substantially greater computational expense than the
single mixture fraction model.
For complex hydrocarbons, like coal, the individual species components are most
often unknown. The empirical model allows you to define the fuel composition
in terms of the ultimate analysis (atomic fractions of C, H, O, N, and S),
along with the lower heating value and heat capacity.
(d) Specify the Empirical Parameters for fuel.
i. Specify the Fuel Lower Calorific Value as 3.53e+07 j/kg.
ii. Specify the value of Fuel Specific Heat as 1000 j/kg-k.
The Fuel Rich Flammability Limit allows you to perform a “partial equilibrium”
calculation, suspending equilibrium calculations when the mixture fraction ex-
ceeds the specified rich limit. This increases the efficiency of the PDF cal-
culation, allowing you to bypass the complex equilibrium calculations in the
fuel-rich region. This is also more physically realistic than the assumption of
full equilibrium. For empirically defined streams, the rich limit is always 1.0
and cannot be altered.
3. Define the boundary species in the Boundary tab.

For an empirically-defined fuel, you need to define the atom fraction of C, H, O,
N, and S elements. The intermediate and product species will be determined auto-
matically.

(a) Under Species Units, select Mole Fraction.

(b) Enter the oxidizer composition.
The oxidizer (air) consists of 21% O2 and 79% N2 by volume.
i. Under Oxid, retain default values for n2 and o2.
(c) Enter the fuel composition.
The fuel composition is entered in mole fractions of the elements C, H, O, and
N. The mole fractions are obtained from the proximate and ultimate analysis
of coal.
The calculation for the proximate and ultimate analysis that yields the elemen-
tal composition of the volatile stream is given in the Appendix: Analysis for
Elemental Composition of Coal.
i. Under Fuel, enter the following values:
Species Mole Fraction
C 0.581
H 0.390
O 0.016
N 0.013
FLUENT will use this information, along with the coal heating value, to define
the species present in the fuel.
Note: All boundary species with a mass or mole fractions of zero will be ig-
nored.
13-10
(d) Under Temperature, enter the inlet temperatures for Fuel as 400 and Oxid as
1500 and click Apply.
The system pressure and inlet stream temperatures are required for the equi-
librium chemistry calculation. The fuel stream inlet temperature for coal com-
bustion should be the temperature at the onset of devolatilization. The oxidizer
inlet temperature should correspond to the air inlet temperature. In this tuto-
rial, the coal devolatilization temperature will be set to 400 K and the air inlet
temperature is 1500 K. The system pressure is one atmosphere.
4. In the Control tab, retain default species to be excluded from the equilibrium cal-
culation.
5. Specify the table parameters and calculate the PDF table.
(a) In the Table tab, retain the default values for all the Table Parameters and
click Apply.
The maximum number of species determines the number of most preponderant
species to consider after the equilibrium calculation is performed.
The minimum temperature should be a few degrees lower than the lowest bound-
ary condition temperature (e.g., the inlet temperature or wall temperature).
In coal combustion systems, the minimum system temperature should also be
set below the temperature at which the volatiles begin to evolve from the coal.
Here, the vaporization temperature at which devolatilization begins will be set
to 400 K. Thus, the minimum system temperature is set to 298 K (the default).

6. Click Calculate PDF Table to compute the non-adiabatic PDF table.

The non-adiabatic calculation requires much more computation than the adiabatic
calculation. FLUENT begins by accessing the thermodynamic data from the database.
The enthalpy field is initialized, the temperature limits are calculated, the stoichio-
metric mixture fractions are calculated, and the enthalpy grid adjusted to account
for the solution parameters.
Mean values of temperature, composition, and density at the discrete mixture-
fraction/mixture-fraction-variance/enthalpy points are then calculated. The result
is a set of tables containing mean values of species mole fractions, density, and
temperature at each discrete value of these three parameters.
7. Click OK to close the Species Model panel.
8. Save the PDF output file (coal.pdf).

File −→ Write −→PDF...
(a) Enter coal.pdf as the Pdf File name.
(b) Click OK to write the file.
By default, the file will be saved as formatted (ASCII, or text). To save a
binary (unformatted) file, turn on the Write Binary Files option in the Select
File panel.
9. Review one slice of the 3D look-up table.

Display −→PDF Tables/Curves...
13-12
(a) Retain default parameters and click Display (Figure 13.3).

The 3D look-up tables are reviewed on a slice-by-slice basis. By default, the
slice selected is that corresponding to the adiabatic enthalpy values. You can
select other slices of constant enthalpy for display, as well.
Z
Y
X
Mean Temperature(K)
FLUENT 6.2 (2d, segregated, pdf20, ske)
Figure 13.3: Non-Adiabatic Temperature Look-Up Table on the Adiabatic Enthalpy Slice
The maximum and minimum values for mean temperature and the corresponding
mean mixture fraction are also reported in the console. The maximum mean tem-
perature is reported as 2782 K at a mean mixture fraction of 0.09.
10. Examine the species/mixture-fraction relationship in the non adiabatic system.

(a) Select Mole Fraction of c<s> as the Plot Variable and click Display (Figure 13.4).
(b) Similarly, plot the mean mole fractions for CO (Figure 13.5).

Z
Y
X
Mole Fraction of c<s>

Figure 13.4: Mole Fractions of C(s)
Z
Y
X
Mole Fraction of co
Figure 13.5: Mole Fractions of CO
13-14
11. Generate a 2D curve of species/mixture-fraction at the adiabatic slice.

(a) Select Mole Fraction of co2 as the Plot Variable.
(b) Select 2D Curve on 3D Surface as the Plot Type and click Display.
1.20e-01
1.00e-01
8.00e-02
Mole 6.00e-02
Fraction
of
co2 4.00e-02
2.00e-02
0.00e+00
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Z
Y Mean Mixture Fraction
X
Mole Fraction of co2

Figure 13.6: Mole Fractions of CO2 (2D Curve)

Step 4: Models: Discrete Phase

The flow of pulverized coal particles will be modeled by FLUENT using the discrete phase
model. This model predicts the trajectories of individual coal particles, each represent-
ing a continuous stream (or mass flow) of coal. Heat, momentum, and mass transfer
between the coal and the gas will be included by alternately computing the discrete phase
trajectories and the gas phase continuum equations.
1. Enable the discrete phase coupling to the continuous phase flow prediction.
Define −→ Models −→Discrete Phase...
13-16
(a) Under Interaction, turn on the Interaction with Continuous Phase option.
This option enables coupling, in which the discrete phase trajectories (along
with heat and mass transfer to the particles) are allowed to impact the gas
phase equations.
If you leave this option turned off, you can track particles but they will have
no impact on the continuous phase flow.
(b) Set the coupling parameter, the Number of Continuous Phase Iterations per
DPM Iteration, to 20.
Use higher values of this parameter in problems that include a high particle
mass loading or a larger grid size. Less frequent trajectory updates can be
beneficial in such problems, in order to converge the gas phase equations more
completely before repeating the trajectory calculation.
(c) Under Tracking Parameters, set the Max. Number of Steps to 10000.
The limit on the number of trajectory time steps is used to abort trajectories
of particles that are trapped in the domain (e.g., in a recirculation).
(d) Retain the default Step Length Factor of 5.
(e) In the Physical Models tab, turn on Particle Radiation Interaction.
(f) Click OK to close the Discrete Phase Model panel.
2. Create the discrete phase coal injections.

The flow of the pulverized coal is defined by the initial conditions that describe the
coal as it enters the gas. FLUENT uses these initial conditions as the starting point
for the time integration of the particle equations of motion (trajectory calculations).
Here, the total mass flow rate of coal (in the half-width of the duct) is 0.1 kg/s
(per unit meter depth). The particles are assumed to obey a Rosin-Rammler size
distribution between 70 and 200 micron diameter.
Define −→ Injections...
(a) Click the Create button in the Injections panel.
This will open the Set Injection Properties panel where you will define the initial
conditions defining the flow of coal particles.

The particle stream will be defined as a group of 10 distinct initial conditions,

all identical except for diameter, which will obey the Rosin-Rammler size dis-
tribution law.
(b) Select group in the Injection Type drop-down list.
(c) Set the Number of Particle Streams to 10.
These inputs tell FLUENT to represent the range of specified initial conditions
by 10 discrete particle streams, each with its own set of discrete initial condi-
tions. Here, this will result in 10 discrete particle diameters, as the diameter
will be varied within the injection group.
(d) Select Combusting under Particle Type.
By selecting Combusting you are activating the submodels for coal devolatiliza-
tion and char burnout. Similarly, selecting Droplet enables the submodels for
droplet evaporation and boiling.
13-18
(e) Select coal-mv in the Material drop-down list.

The Material list contains the combusting particle materials in the FLUENT
database. You can select an appropriate coal from this list and review or
modify its properties in the Materials panel (see Step 5-2: Materials).
(f) Select rosin-rammler in the Diameter Distribution drop-down list.
The coal particles have a nonuniform size distribution with diameters ranging
from 70 µm to 200 µm. The size distribution fits the Rosin-Rammler equation,
with a mean diameter of 134 µm and a spread parameter of 4.52.
(g) Retain o2 in the Oxidizing Species drop-down list.
(h) Specify the range of initial conditions under Point Properties starting with the
following values for First Point:
• X-Position: 0.001 m
• Y-Position: 0.03124 m
• X-Velocity: 10 m/s
• Y-Velocity: 5 m/s
• Temperature: 300 K
• Total Flow Rate: 0.1 kg/s
• Min. Diameter: 70e-6 m
• Max. Diameter: 200e-6 m
• Mean Diameter: 134e-6 m
• Spread Parameter: 4.52
(i) Under Last Point, specify identical values for position, velocity, and tempera-
ture.
(j) Click on the Turbulent Dispersion tab to define the turbulent dispersion.
The panel will change to show the related inputs.

i. Under Stochastic Tracking, turn on Discrete Random Walk Model.

Stochastic tracks model the effect of turbulence in the gas phase on the
particle trajectories. Including stochastic tracking is important in coal
combustion simulations, to simulate realistic particle dispersion.
ii. Set the Number of Tries to 10 and click OK.
The new injection (named injection-0, by default) appears in the Injections
panel.
13-20
The Injections panel can be used to copy and delete injection definitions. You
list the initial conditions of particle streams defined by an injection in the
console window.
The listing for the injection-0 group will show 10 particle streams, each with a
unique diameter between the specified minimum and maximum value, obtained
from the Rosin-Rammler distribution, and a unique mass flow rate.

Step 5: Materials
1. Specify the continuous phase (pdf-mixture) material.

All thermodynamic data for the continuous phase, including density, specific heat,
and formation enthalpies are extracted from the chemical database when the non-
premixed combustion model is used. These properties are transferred as the pdf-
mixture material, for which only transport properties, such as viscosity and thermal
conductivity, need to be defined.
(a) Set Thermal Conductivity to 0.025 (constant).

(b) Set Viscosity to 2e-5 (constant).
(c) Select wsggm-cell-based in the drop-down list for the Absorption Coefficient.
This specifies a composition-dependent absorption coefficient, using the weighted-
sum-of-gray-gases model.
See Section 12.3.9 of the User’s Guide for more details.
13-22
(d) Click the Change/Create button.

You can click on the View... button next to Mixture Species to view the species
included in the pdf-mixture material. These are the species included during the
system chemistry setup. The Density and Cp laws cannot be altered: these properties
are stored in the non-premixed combustion look-up tables.
FLUENT uses the gas law to compute the mixture density and a mass-weighted
mixing law to compute the mixture cp . When the non-premixed combustion model
is used, do not alter the properties of the individual species. This will create an
inconsistency with the PDF look-up table.
2. Define the discrete phase material.

(a) In the Material Type drop-down list, select combusting-particle.
The combusting-particle material type appears because you have activated com-
busting particles using the Set Injection Properties panel. Other discrete phase
material types (droplets, inert particles) will appear in this list if you have
created injections of those types.
(b) Retain the default selection (coal-mv) in the Fluent Combusting Particle Mate-
rials list.

This is the combusting particle material type that you selected from the list of
database options in the Set Injection Properties panel. Additional combusting
particle materials can be copied from the property database, if required. Click
the Fluent Database... button to view the combusting-particle materials that
are available. Modify the property settings for the selected material, coal-mv.
How FLUENTuses these inputs are defined in the Appendix: Discrete Phase
Material Properties.
(c) Modify the coal-mv material properties as per following constant values:
Properties Values
Density kg/m3 1300
Cp J/kg-K 1000
Thermal Conductivity w/m-k 0.0454
Latent Heat 0
Vaporization Temperature K 400
Volatile Component Fraction (%) 28
Binary Diffusivity w/m-k 5e-4
Particle Emissivity 0.9
Particle Scattering Factor 0.6
Swelling Coefficient 2
Burnout Stoichiometric Ratio 2.67
Combustible Fraction (%) 64
Note: The values for the Vaporization Temperature should be consistent with
the fuel temperature considered in Step 3-3: Non Adiabatic PDF Table. The
Volatile Component Fraction and Combustible Fraction should be consistent with
the volatiles and char ratios in the proximate analysis of the coal shown in the
Appendix: Analysis for Elemental Composition of Coal.
(d) For Devolatilization Model, select single rate in the drop-down list and accept
default values in the Single Rate Devolatilization Model panel.
(e) For the Combustion Model, select kinetics/diffusion limited and accept default
values in the Kinetics/Diffusion-Limited Combustion Model panel.
13-24
(f) Click Change/Create and close the Materials panel.


1. Keep the default operating conditions.
13-26

Hint: You can click the mouse probe button (right button, by default) on the required
boundary zone in the graphics display window to select that zone in the Boundary
Conditions panel.
1. Set the following conditions for the zone, velocity-inlet-2 (low-speed inlet boundary).
Properties Values
Velocity Specification Method Magnitude and Direction
Velocity Magnitude(m/s) 15
Temperature(k) 1500
Turbulent Specification Method Intensity and Hydraulic Diameter
Turbulence Intensity(%) 10
Hydraulic Diameter(m) 0.75

Turbulence parameters are defined based on intensity and hydraulic diameter. The
relatively large turbulence intensity of 10% may be typical for combustion air flows.
The hydraulic diameter has been set to twice the height of the 2D inlet stream.
For the non-premixed combustion calculation, you have to define the inlet Mean
Mixture Fraction and Mixture Fraction Variance. For coal combustion, all fuel comes
from the discrete phase and thus the gas phase inlets have zero mixture fraction.
Therefore, you can accept the zero default settings.
2. Set the following conditions for the velocity-inlet-8 zone (high-speed inlet boundary).
Properties Values
Velocity Specification Method Magnitude and Direction
Velocity Magnitude(m/s) 50
Temperature(k) 1500
Hydraulic Diameter(m) 0.25
13-28
3. Set the following conditions for the pressure-outlet-6 zone (exit boundary).
Properties Values
Backflow Total Temperature 2000
Hydraulic Diameter(m) 1
The exit gauge pressure of zero defines the system pressure at the exit to be the
operating pressure. The backflow conditions for scalars (temperature, mixture frac-
tion, turbulence parameters) will be used only if flow is entrained into the domain
through the exit. It is a good idea to use reasonable values in case flow reversal
occurs at the exit at some point during the solution process.

4. Set conditions for the wall-7 zone (furnace wall).

The furnace wall will be treated as an isothermal boundary with a temperature of
1200 K.
(a) Under Thermal Conditions, select Temperature.

(b) Enter 1200 in the Temperature field.
Note: The default boundary condition for particles that hit the wall is reflect, as
shown under DPM. Alternate treatments can be selected, using the BC Type
list, for particles that hit the wall.
13-30
Step 8: Solution
1. Set the P1 under-relaxation factor to 1.
2. Initialize the flow field using conditions at velocity-inlet-2.

(a) Select velocity-inlet-2 in the Compute From list.

(b) Click the Init button to initialize the flow field, and then close the panel.
Use the Init button to initialize the flow field data.The Apply button does
! not initialize the flow field data. It only allows you to store your initializa-
tion parameters for later use.
Note: Here, with very high pre-heat of the oxidizer stream, you can start the com-
bustion calculation from the inlet-based initialization. In general, you may
need to start your coal combustion calculations by patching a high-temperature
region and performing a discrete phase trajectory calculation. This provides the
initial volatile and char release required to initiate combustion. The Solve/Initialize/
Patch... menu item and the solve/dpm-update text command can be used to
perform this initialization.
3. Enable the display of residuals during the solution process.

(a) Enable Plot under Options and then click OK.

4. Save the case file (coal.cas).

5. Begin the calculation by requesting 400 iterations.

Note: The default convergence criteria will be met in about 140 iterations.
6. Save the converged flow data (coal.dat).

13-32
1. Display the predicted temperature field (Figure 13.7).
Hint: Use the Views panel (Display/Views...) to mirror the display about the sym-
metry plane.
The peak temperature in the system is about 2280 K.
2. Display the devolatilization rate (Figure 13.8).

(a) Select Discrete Phase Model... and DPM Evaporation/Devolatilization in the
Contours of drop-down lists.
Figure 13.8 shows that volatiles are released after the coal travels about one eighth of
the furnace length. The onset of devolatilization occurs when the coal temperature
reaches the specified value of 400 K.
3. Display the char burnout rate (Figure 13.9) by selecting DPM Burnout from the
lower drop-down list.
The char burnout occurs after complete devolatilization. Figure 13.9 shows that
burnout is complete at about three-quarters of the furnace.

2.25e+03
2.20e+03
2.15e+03
2.09e+03
2.04e+03
1.99e+03
1.94e+03
1.88e+03
1.83e+03
1.78e+03
1.73e+03
1.67e+03
1.62e+03
1.57e+03
1.52e+03
1.46e+03
1.41e+03
1.36e+03
1.31e+03 Y
1.25e+03 Z X
1.20e+03

Figure 13.7: Temperature Contours
2.54e-03
2.41e-03
2.29e-03
2.16e-03
2.03e-03
1.91e-03
1.78e-03
1.65e-03
1.52e-03
1.40e-03
1.27e-03
1.14e-03
1.02e-03
8.89e-04
7.62e-04
6.35e-04
5.08e-04
3.81e-04
2.54e-04 Y
1.27e-04 Z X
0.00e+00
Contours of DPM Evaporation/Devolatilization (kg/s)

Figure 13.8: Devolatilization Rate
13-34
3.89e-04
3.70e-04
3.51e-04
3.31e-04
3.12e-04
2.92e-04
2.73e-04
2.53e-04
2.34e-04
2.14e-04
1.95e-04
1.75e-04
1.56e-04
1.36e-04
1.17e-04
9.74e-05
7.79e-05
5.84e-05
3.89e-05 Y
1.95e-05 Z X
0.00e+00
Contours of DPM Burnout (kg/s)

Figure 13.9: Char Burnout Rate

4. Display the Mean Mixture Fraction distribution (Figure 13.10).

(a) Select Pdf... and Mean Mixture Fraction in the Contours of drop-down lists.
The mixture-fraction distribution shows where the char and volatiles released from
the coal exist in the gas phase.
3.71e-02
3.52e-02
3.34e-02
3.15e-02
2.96e-02
2.78e-02
2.59e-02
2.41e-02
2.22e-02
2.04e-02
1.85e-02
1.67e-02
1.48e-02
1.30e-02
1.11e-02
9.26e-03
7.41e-03
5.56e-03
3.71e-03 Y
1.85e-03 Z X
0.00e+00
Contours of Mean Mixture Fraction

Figure 13.10: Mixture-Fraction Distribution
5. Display the oxygen distribution (Figure 13.11).

(a) Select Species... and Mass Fraction of o2 in the Contours of drop-down lists.
Note: Although transport equations are solved only for the mixture fraction and its
variance, you can display the chemical species concentrations as predicted by
the PDF equilibrium chemistry model.
6. Similarly, display mass fraction distributions for other species — CO2 (Figure 13.12),
H2 O (Figure 13.13), CO (Figure 13.14).
13-36
2.33e-01
2.27e-01
2.22e-01
2.16e-01
2.11e-01
2.05e-01
2.00e-01
1.94e-01
1.89e-01
1.83e-01
1.78e-01
1.72e-01
1.67e-01
1.61e-01
1.55e-01
1.50e-01
1.44e-01
1.39e-01
1.33e-01 Y
1.28e-01 Z X
1.22e-01
Contours of Mass fraction of o2

Figure 13.11: O2 Distribution
1.21e-01
1.15e-01
1.09e-01
1.03e-01
9.67e-02
9.07e-02
8.46e-02
7.86e-02
7.25e-02
6.65e-02
6.04e-02
5.44e-02
4.84e-02
4.23e-02
3.63e-02
3.02e-02
2.42e-02
1.81e-02
1.21e-02 Y
6.04e-03 Z X
0.00e+00
Contours of Mass fraction of co2

Figure 13.12: CO2 Distribution

1.64e-02
1.56e-02
1.48e-02
1.40e-02
1.31e-02
1.23e-02
1.15e-02
1.07e-02
9.85e-03
9.03e-03
8.21e-03
7.39e-03
6.57e-03
5.75e-03
4.93e-03
4.10e-03
3.28e-03
2.46e-03
1.64e-03 Y
8.21e-04 Z X
0.00e+00
Contours of Mass fraction of h2o

Figure 13.13: H2 O Distribution
4.54e-03
4.31e-03
4.08e-03
3.86e-03
3.63e-03
3.40e-03
3.18e-03
2.95e-03
2.72e-03
2.50e-03
2.27e-03
2.04e-03
1.81e-03
1.59e-03
1.36e-03
1.13e-03
9.07e-04
6.81e-04
4.54e-04 Y
2.27e-04 Z X
0.00e+00
Contours of Mass fraction of co

Figure 13.14: CO Distribution
13-38
7. Display the particle trajectory of one particle stream (Figure 13.15).

Display −→Particle Tracks...
(a) Retain Particle Variables... and Particle Residence Time under Color by.
(b) Select injection-0 in the Release from Injections list.
(c) Turn on Track Single Particle Stream, set the Stream ID to 5 and click Display.

4.21e-01
4.00e-01
3.79e-01
3.58e-01
3.37e-01
3.16e-01
2.95e-01
2.74e-01
2.53e-01
2.32e-01
2.11e-01
1.90e-01
1.69e-01
1.47e-01
1.26e-01
1.05e-01
8.43e-02
6.32e-02
4.21e-02 Y
2.11e-02 Z X
0.00e+00
Particle Traces Colored by Particle Residence Time (s)

Figure 13.15: Trajectories of Particle Stream 5 Colored by Particle Residence Time
13-40
Step 10: Energy Balances and Particle Reporting

FLUENT can report the overall energy balance and details of the heat and mass transfer
from the discrete phase.
1. Compute the fluxes of heat through the domain boundaries.

(a) Select Total Heat Transfer Rate under Options.

(b) Under Boundaries, select pressure-outlet-6, velocity-inlet-2, velocity-inlet-8, and
wall-7 zones.
(c) Click Compute.
Positive flux reports indicate heat addition to the domain. Negative values indicate
heat leaving the domain. In reacting flows, the heat report uses total enthalpy
(sensible heat plus heat of formation of the chemical species). The net “imbalance”
of total enthalpy (about 11.26 kW) represents the total enthalpy addition from the
discrete phase.
2. Compute the volume sources of heat transferred between the gas and discrete par-
ticle phase.
Report −→Volume Integrals...

(a) Select Sum under Report Type.

(b) Select Discrete Phase Model... and DPM Enthalpy Source in the Field Variable
drop-down lists.
(c) Select fluid-1 under Cell Zones and click Compute.
The total enthalpy transfer from gas to the discrete phase is about -11.05 kW, as
expected based on the boundary flux report above. This represents the total en-
thalpy addition from the discrete phase to the gas during devolatilization and char
combustion processes.
3. Obtain a summary report on the particle trajectories.

The discrete phase model summary report provides detailed information about the
particle residence time, heat and mass transfer between the continuous and discrete
phases, and (for combusting particles) char conversion and volatile yield.
(a) Select Summary under Report Type.
(b) Select injection-0 and click Track.
FLUENT will report the summary in the console window. You can write the
report to a file by selecting File under Report to.
(c) Review the summary printed in the console window:
13-42
DPM Iteration ....

number tracked=100, escaped=100, aborted=0, trapped=0, evaporated=0, incomplete=0
Fate Number
Elapsed Time (s) Injection,
Min Max Avg Std Dev Min Max
---- ----- -------- --------- --------- ---------- -------- --------
Escaped-Zone6 100 2.511e-01 4.793e-01 3.31e-01 5.248e-02 inj..n-0 1 inj..n-0 7
(*)- Mass Transfer Summary -(*)
Fate Mass Flow (kg/s)

Initial Final Change
---- ---------- ---------- ----------
Escaped - Zone 6 1.000e-01 8.003e-03 -9.200e-02
(*)- Energy Transfer Summary -(*)
Fate Heat Content (W)

Initial Final Change
---- ---------- ---------- ----------
Escaped - Zone 6 -1.463e+03 9.808e+03 1.127e+04
(*)- Combusting Particles -(*)
Fate Volatile Content (kg/s) Char Content (kg/s)

Initial Final %Conv Initial Final %Conv
---- ---------- ---------- ------- ---------- ---------- -------
Escaped - Zone 6 2.800e-02 0.000e+00 100.00 6.400e-02 3.803e-06 99.99
The report shows that the average residence time of the coal particles is about 0.33
seconds. Volatiles are completely released within the domain and the char conver-
sion is 99.99% .

Extra: You can obtain a detailed report of the particle position, velocity, diameter, and
temperature along the trajectories of individual particles. A detailed track reporting
is useful if you are trying to understand unusual or important details in the discrete
model behavior. To generate the report, visit the Particle Tracks panel and do the
following:
1. Select Step By Step under Report Type.
2. Select File under Report to.
3. Enable the Track Single Particle Stream option, and set the Stream ID to the
desired particle stream.
4. Click Write... to bring up the Select File dialog box and enter the name of the
file to be written.
This file can be viewed with a text editor.
Summary
Coal combustion modeling involves the prediction of volatile evolution and char burnout
from pulverized coal along with simulation of the combustion chemistry occurring in the
gas phase. In this tutorial you learned how to use the non-premixed combustion model to
represent the gas phase combustion chemistry. In this approach the fuel composition was
defined empirically and the fuel was assumed to react according to the equilibrium system
data. This equilibrium chemistry model can be applied to other turbulent, diffusion-
reaction systems. You can also model coal combustion using the finite-rate chemistry
model.
You also learned how to set up and solve a problem involving a discrete phase of com-
busting particles. You created discrete phase injections, activated coupling to the gas
phase, and defined the discrete phase material properties. These procedures can be used
to set up other simulations involving reacting or inert particles.
13-44
Appendix
Coal Analysis for Elemental Composition
From proximate analysis, the fuel considered here consists of 28% volatiles, 64% char,
and 8% ash. Use this information, along with the ultimate analysis, to define the coal
composition. Define the composition of the fuel stream (char) to be 100% C(S). The fuel
stream composition (char and volatiles) is derived as follows.
Begin by converting the proximate data to a dry-ash-free basis:
Proximate Analysis Wt % Wt %
(dry) (DAF)
Volatiles 28 30.4
Char (C(s)) 64 69.6
Ash 8 -
For the dry-ash-free coal, the ultimate analysis is:
Element Wt % (DAF)
C 89.3
H 5.0
O 3.4
N 1.5
S 0.8
For simplifying modeling, the sulfur content of the coal can be combined into the nitrogen
mass fraction, to yield:
Element Wt % (DAF)
C 89.3
H 5.0
O 3.4
N 2.3
S -
Combine the proximate and ultimate analysis data to yield the following elemental com-
position of the volatile stream:
Element Wt % Moles Mole Fraction

C 89.3 7.44 0.581
H 5.0 5 0.390
O 3.4 0.21 0.016
N 2.3 0.16 0.013
Total 12.81

Discrete Phase Material Properties

FLUENT uses the material properties for the discrete phase as follows:
• Density impacts the particle inertia and body forces (when the gravitational accel-
eration is non-zero).
• Cp determines the heat required to change the particle temperature.
• Latent Heat is the heat required to vaporize the volatiles. This can usually be set
to zero when the non-premixed combustion model is used for coal combustion. If
the volatile composition has been selected in order to preserve the heating value of
the fuel, the latent heat has been effectively included. (You would, however, use a
non-zero latent heat if water content had been included in the volatile definition as
vapor phase H2 O.)
• Vaporization Temperature is the temperature at which the coal devolatilization be-

gins. It should be set equal to the fuel inlet temperature used in PDF table.
• Volatile Component Fraction determines the mass of each coal particle that is de-
volatilized.
• Binary Diffusivity is the diffusivity of oxidant to the particle surface and is used in
the diffusion-limited char burnout rate.
• Particle Emissivity is the emissivity of the particles. It is used to compute radiation

heat transfer to the particles.
• Particle Scattering Factor is the scattering factor due to particles.
• Swelling Coefficient determines the change in diameter during coal devolatilization.

A swelling coefficient of 2 implies that the particle size will double as the volatile
fraction is released.
• Burnout Stoichiometric Ratio is used in the calculation of the diffusion-controlled

burnout rate. Otherwise, this parameter has no impact when the non-premixed
combustion model is used. When finite-rate chemistry is used instead, the stoichio-
metric ratio defines the mass of oxidant required per mass of char. The default
value represents oxidation of C(s) to CO2 .
• Combustible Fraction is the mass fraction of char in the coal particle. It determines
the mass of each coal particle that is consumed by the char burnout submodel.
13-46
Tutorial 14. Modeling Surface Chemistry
Introduction
In chemically reacting laminar flows, such as those encountered in chemical vapor de-
position (CVD) applications, accurate modeling of time-dependent hydrodynamics, heat
and mass transfer, and chemical reactions (including wall surface reactions) is important.
Tutorials 12 and 13 deal with reacting flows with applications in gaseous fuel and coal
combustion.
In this tutorial, surface reactions are considered. Modeling the reactions taking place at
gas-solid interfaces is complex and involves several elementary physico-chemical processes
like adsorption of gas-phase species on the surface, chemical reactions occurring on the
surface, and desorption of gases from the surface back to the gas phase.
In this tutorial, you will learn how to:
• Create new materials and set the mixture properties
• Model surface reactions involving site species
• Enable physical models and define boundary conditions for a chemically reacting
laminar flow involving wall surface reactions
• Calculate the deposition solution using the segregated solver
• Examine the flow results using graphics
Prerequisites
Before beginning, you should read Chapter 13 of the User’s Guide for more information
about species transport, chemically reacting flows, wall surface reaction modeling, and
chemical vapor deposition.. In particular, you should be familiar with the Arrhenius rate
equation, as this equation is used for the surface reactions modeled in this tutorial.

Modeling Surface Chemistry
Inlet
Rotating
Disk
Outlet
Figure 14.1: An Outline of the Reactor Configuration
Problem Description
A rotating disk CVD reactor for the growth of Gallium Arsenide (GaAs) shown in Fig-
ure 14.1 will be modeled.
The process gases, Trimethyl Gallium (Ga(CH3 )3 ) and Arsine (AsH3 ) enter the reactor
at 293 K through the inlet at the top. These gases flow over the hot, spinning disk
depositing thin layers of gallium and arsenide on it in a uniform, repeatable manner.
The disk rotation generates a radially pumping effect, which forces the gases to flow in a
laminar manner down to the growth surface, outward across the disk, and finally to be
discharged from the reactor.
The semiconductor materials Ga(s) and As(s) are deposited on the heated surface gov-
erned by the following surface reactions.
AsH3 + Ga s → Ga + As s + 1.5H2 (14.1)

Ga(CH3 )3 + As s → As + Ga s + 3CH3 (14.2)
As mentioned earlier, the inlet gas is a mixture of trimethyl gallium and arsine. In the
inlet mixture the mass fraction of Ga(CH3 )3 is 0.15 and AsH3 is 0.4. The mixture velocity
at the inlet is 0.02189 m/s. The disk rotates at 80 rad/sec, and the top wall (wall-1) is
heated to 473 K, and the sidewalls (wall-2) of the reactor are maintained at 343 K. The
susceptor (wall-4) is heated to a uniform temperature of 1023 K, and the bottom wall
(wall-6) is at 303 K. These CVD reactors are typically known as cold-wall reactors, where
only the wafer surface is heated to higher temperatures, while the remaining reactor walls
are maintained at low temperatures.
In this tutorial, simultaneous deposition of Ga and As is simulated and examined.
The mixture properties and the mass diffusivity are determined based on kinetic the-
14-2
ory. Detailed surface reactions with multiple sites and site species, and full multi-
component/thermal diffusion effects are also included in the simulation.
The purpose of this tutorial is to demonstrate surface reaction capabilities in FLUENT.
Convective heat transfer is considered to be the dominant mechanism compared to ra-
diative heat transfer, thus radiation effects are ignored.
Setup and Solution

Preparation
1. Download surface_chem.zip from the Fluent Inc. User Services Center or copy
Tutorial 1).
2. Unzip surface_chem.zip.
surface.msh can be found in the /surface chem folder created after unzipping the
file.
3. Start the 3ddp version of FLUENT.

Step 1: Grid
1. Read in the mesh file surface.msh.
2. Check the grid.

Grid −→Check
Note: The grid check lists the minimum and maximum x and y values from the
grid, and reports on a number of other grid features that are checked. Any
errors in the grid would be reported at this time. For instance, the cell volumes
must never be negative. Note that the domain extents are reported in units of
meters, the default unit of length in FLUENT. Since this grid was created in
units of centimeters, the Scale Grid panel will be used to scale the grid into
meters.
3. Scale the grid.

Grid −→Scale...
(a) In the Units Conversion drop-down list, select cm to complete the phrase Grid
Was Created In cm (centimeters).
The final Domain Extents should appear as in the panel above.
Note: Because the default SI units will be used in this tutorial, there is no need to
change any units.
14-4

Extra: You can use the left mouse button to rotate the image and view it from
different angles. You can use the right mouse button to check which zone
number corresponds to each boundary. If you click the right mouse button
on one of the boundaries in the graphics window, its name and type will be
them quickly. Use the middle mouse button to zoom the image.

Figure 14.2: Grid Display
14-6
Step 2: Models
In this problem, the energy equation and the species conservation equations will be solved,
along with the momentum and continuity equations.




(a) Under Model, select Species Transport.

This will expand the Species Model panel.
(b) Under Reactions, select Volumetric and Wall Surface.
(c) Under Wall Surface Reaction Options, select Mass Deposition Source.
Mass Deposition Source is selected because there is a certain loss of mass due
to the surface deposition reaction, i.e., As(s) and Ga(s) are being deposited
out. If you were to do an overall mass balance without taking this fact into
account, you would end up with a slight imbalance.
(d) Keep the Diffusion Energy Source option turned on.
14-8
Note: This includes the effect of enthalpy transport due to species diffusion
in the energy equation, which contributes to the energy balance, especially
for the case of Lewis numbers far from unity.
(e) Select Full Multicomponent Diffusion and Thermal Diffusion.
Note: The Full Multicomponent Diffusion activates Stefan-Maxwell’s equa-
tions and computes the diffusive fluxes of all species in the mixture to
all concentration gradients. The Thermal Diffusion effects cause heavy
molecules to diffuse less rapidly, and light molecules to diffuse more rapidly,
towards heated surfaces.
(f) Click OK.
The console window will list the properties that are required for the models
that you have enabled. You will see an Information dialog box, reminding you
to confirm the property values that have been extracted from the database.
(g) Click OK in the Information dialog box to continue.

Step 3: Materials
1. Create the gas-phase species (AsH3 , Ga(CH3 )3 , CH3 , H2 ), the site species (Ga s
and As s), and solid species (Ga and As).
(a) Create species AsH3 .
i. In the Materials panel, select fluid under Material Type.
ii. Select nitrogen under Fluid Materials to create the new material.
iii. Set the Mixture to none.
iv. Enter arsine under Name.
v. Enter ash3 under Chemical Formula.
vi. Specify the following for each of the properties:
14-10
Parameter Value
Cp kinetic-theory
Thermal Conductivity kinetic-theory
Viscosity kinetic-theory
Molecular Weight 77.95
Standard State Enthalpy 0
Standard State Entropy 130579.1
Reference Temperature 298.15
L-J Characteristic Length 4.145
L-J Energy Parameter 259.8
Degrees of Freedom 0
Ignore the Density parameter as the density will be set to incompressible-
ideal-gas-law for mixture.
vii. Click Change/Create to create the new material.
viii. FLUENT will ask if you would like to overwrite nitrogen. Click No in the
Question panel.

(b) Create the other species following the same procedure as for AsH3 .
i. The parameter values for each of the species is as per the table given
below:
Parameter Ga(CH3 )3 CH3 H2 Ga s As s Ga As

Name tmg ch3g hydrogen ga s as s ga as
Chemical For- ga<ch3>3 ch3 h2 ga s as s ga as
mula
Cp kinetic- kinetic- kinetic- 520.64 520.64 1006.43 1006.43
theory theory theory
Thermal Con- kinetic- kinetic- kinetic- 0.0158 0.0158 kinetic- kinetic-
ductivity theory theory theory theory theory
Viscosity kinetic- kinetic- kinetic- 2.125 2.125 kinetic- kinetic-
theory theory theory e-05 e-05 theory theory
Molecular 114.83 15 2.02 69.72 74.92 69.72 74.92
Weight
Standard State 0 2.044 0 -3117.71 -3117.71 0 0
Enthalpy e+07
Standard State 130579.1 257367.6 130579.1 154719.3 154719.3 0 0
Entropy
Reference 298.15 298.15 298.15 298.15 298.15 298.15 298.15
Temperature
L-J Character- 5.68 3.758 2.827 - - 0 0
istic Length
L-J Energy Pa- 398 148.6 59.7 - - 0 0
rameter
Degrees of 0 0 5 - - - -
Freedom
ii. Click Change/Create to create the new material.

iii. Click No in the Question panel.
To enter complex formulae such as Ga(CH3 )3 in the text entry box, ‘<’ and
‘>’ are used instead of ‘(’ and ‘)’, respectively.
2. Set the mixture species.

(a) Under Material Type, select mixture.
(b) Enter gaas deposition under Name.
(c) Click Change/Create.
(d) To overwrite the mixture-template, click Yes in the Question panel.
14-12
(e) Under Properties, click Edit... to the right of Mixture Species.

This will open the Species panel. Here you will set the Selected Species, Selected
Site Species, and Selected Solid Species from the Available Materials list using
the Add and Remove buttons.
The species under each species type are:
Selected Species Selected Site Species Selected Solid Species
ash3 ga s ga
ga<ch3>3 as s as
ch3 - -
h2 - -
The species should appear in the same order as shown in the above table.
i
(f) To set the species follow the procedure listed below:
i. To remove an unwanted species from the Selected Species list, select the
species and click Remove under Selected Species.
ii. Select the required species in the Available Materials list.
iii. Click Add under the corresponding species list.
iv. Click OK after all the species are set under the respective categories.

3. Set the mixture reactions.

(a) In the Materials panel, under Properties, click Edit... to the right of Reaction.
This will open the Reactions panel.
i. Increase the Total Number of Reactions to 2, and define the following
reactions:
Ga(CH3 )3 + As s → As + Ga s + 3CH3 (14.3)
AsH3 + Ga s → Ga + As s + 1.5H2 (14.4)
CH3 further reacts with H (3CH3 + 1.5H2 → 3CH4 ) on the substrate

producing CH4 .
The parameters for the equations are as follows:
Parameter For Equation 14.3 For Equation 14.4

Reaction Name gallium-dep arsenic-dep
Reaction ID 1 2
Reaction Type Wall Surface Wall Surface
Number of Reactants 2 2
Species ash3, ga s ga<ch3>3, as s
Stoich. Coefficient ash3=1, ga s=1 ga<ch3>3=1, as s=1
Rate Exponent ash3=1, ga s=1 ga<ch3>3=1, as s=1
Arrhenius Rate PEF=1e+06, AE=0, PEF=1e+12, AE=0,
TE=0.5 TE=0.5
Number of Products 3 3
Species ga, as s, h2 as, ga s, ch3
Stoich. Coefficient ga=1, as s=1, h2=1.5 as=1, ga s=1, ch3=3
Rate Exponent as s=0, h2=0 ga s=0, ch3=0
Here, PEF = Pre-Exponential Factor, AE = Activation Energy, and TE =

Temperature Exponent.
14-14
ii. Change the ID to 2 and set the parameters for the second equation as
shown in the above table.
iii. Click OK to save your data and close the panel.
4. Set the reaction mechanisms for the mixture.

(a) In the Materials panel, under Properties, click Edit... to the right of Mechanism.
This will open the Reaction Mechanisms panel.
(b) Enter gaas-ald under Name.
(c) Retain the Number of Mechanisms as 1.
(d) Set the Reaction Type to Wall Surface.
(e) Under Reactions, select gallium-dep and arsenic-dep.
(f) Increase the Number of Sites to 1.
(g) Set the Site Density as 1e-08.

(h) Click Define... to the right of site-1.

This opens the Site Parameters panel.
i. In the Site Parameters panel, set the Total Number of Site Species to 2.
ii. Select ga s as the first site species and set the Site Coverage to 0.7.
iii. Select as s as the second site species and set the Site Coverage to 0.3.
(i) Click Apply and Close the panel.

(j) Click OK in the Reaction Mechanisms panel to close the panel.
14-16
5. In the Materials panel, set Density to incompressible-ideal-gas.
6. Set Cp to mixing-law.
7. Set Thermal Conductivity to mass-weighted-mixing-law.
8. Set Viscosity to mass-weighted-mixing-law.
9. Set Mass Diffusivity to kinetic-theory.
10. Set Thermal Diffusion Coefficient to kinetic-theory.
11. Click Change/Create and close the Materials panel.


1. Set the Operating Pressure to 10000 pascals.
2. Enable Gravity.
3. Set the Gravitational Acceleration in the Z direction as 9.81.
4. Set the Operating Temperature to 303 K.
5. Click OK to close the panel.
14-18

1. Keep the default settings for outflow.
2. Set the conditions for velocity-inlet.

(a) Retain the default Velocity Specification Method as Magnitude,Normal to Bound-
ary.
(b) Retain the default Reference Frame as Absolute.
(c) Set the Velocity Magnitude to 0.02189 m/s
(d) Set the Temperature to 293 K.
(e) Set the Species Mass Fractions for ash3 as 0.4, ga<ch3>3 as 0.15, and ch3 as
0.

(f) Click OK to close the panel.

(a) Under Thermal Conditions, select Temperature and set the Temperature to
473 K.
(b) Click OK to close the panel.
4. Similarly, set the boundary conditions for wall-2.

343 K.

1023 K.
(b) In the Momentum section of the panel, set Wall Motion to Moving Wall, Motion
to Absolute and Rotational and set the Speed to 80 rad/s.
(c) Retain other defaults.
14-20
(d) Under the Species section of the panel, enable Reaction and set Mechanisms to
gaas-ald.

(e) Click OK to close the panel.

720 K.
(b) In the Momentum section of the panel, set Wall Motion to Moving Wall, Motion
to Absolute and Rotational and set the Speed to 80 rad/s.
(c) Retain other defaults.
(d) Click OK to close the panel.

303 K.
8. Use the TUI commands to turn off diffusion at the inlet. In the console window,
type the commands shown in boxes in the dialog below.
Hint: You may need to enter press the <Enter> key to get the > prompt.
> define/models/species/inlet-diffusion?
Include diffusion at inlets? [yes] no
14-22
Step 6: Solution
(a) Change the Under-Relaxation Factor as follows:
Parameter URF Parameter URF
Pressure 0.1 ash3 1
Density 0.3 ga<ch3>3 1
Body Forces 1 ch3 1
Momentum 0.2 Energy 0.9
Hint: You will need to scroll down the Under-Relaxation Factors list to see the
species and Energy.
(b) Under Discretization, retain the default First Order Upwind for Momentum, all
the species and Energy.

2. Initialize the flow field using the boundary conditions set at velocity-inlet.
(a) Select velocity-inlet in the Compute From drop-down list.

14-24

(a) Select Plot under Options.

(b) Set the Convergence Criterion for Continuity to 1e-05.
(c) Click OK to close the panel.
4. Save the case file (surface.cas).

Residuals
continuity
x-velocity
1e+01
y-velocity
z-velocity
energy 1e+00
ash3
gach3 1e-01
ch3
1e-02
1e-03
1e-04
1e-05
1e-06
1e-07
1e-08
1e-09
0 200 400 600 800 1000 1200 1400 1600 1800 2000
Y
Z X Iterations
Scaled Residuals
FLUENT 6.2 (3d, dp, segregated, spe, lam)
Figure 14.3: Scaled Residuals
6. Save the case and data files (surface.cas and surface.dat).

14-26
1. Create an iso-surface near wall-4.
(a) In the Iso-Surface panel, select Grid and Z-Coordinate under Surface of Constant.
(b) Click Compute.
(c) Enter 0.075438 under Iso-Values.
(d) Enter z=0.07 under New Surface Name.
(e) Click Create.
2. Display contours of temperature on the plane surface. (Figure 14.4).

(a) Select Temperature... and Static Temperature in the Contours Of drop-down
list.
(b) Enable Filled under Options.
(c) Select z=0.07 under Surfaces.
(d) Click Display.

1.02e+03
9.86e+02
9.50e+02
9.14e+02
8.77e+02
8.40e+02
8.04e+02
7.68e+02
7.31e+02
6.94e+02
6.58e+02
6.22e+02
5.85e+02
5.48e+02
5.12e+02
4.76e+02
4.39e+02
4.02e+02
3.66e+02
3.30e+02 Z X
2.93e+02
Y

Figure 14.4: Temperature Contours near wall-4
The temperature contours shows the temperature distribution across a plane just
above the rotating disk. You can see that the disk has a temperature of 1023 K.
14-28
3. Display contours of surface deposition rates of ga.(Figure 14.5).

(a) Select Species... and Surface Deposition Rate of ga in the Contours Of drop-
down list.
(b) Select wall-4 under Surfaces.
(c) Click Display.
You may need to use the left mouse button to rotate the image so that you can
see the contours on the top side of wall-4 where the deposition takes place.
3.89e-05
3.70e-05
3.50e-05
3.31e-05
3.11e-05
2.92e-05
2.72e-05
2.53e-05
2.34e-05
2.14e-05
1.95e-05
1.75e-05
1.56e-05
1.36e-05
1.17e-05
9.73e-06
7.78e-06
5.84e-06
3.89e-06
1.95e-06 Z X
0.00e+00
Y
Contours of Surface Deposition Rate of ga (kg/m2-s)

Figure 14.5: Contours of Surface Deposition Rate of ga
Figure 14.5 shows the gradient of surface deposition rate of ga. The maximum
deposition is seen at the center of the disk.
4. Display contours of surface coverage of ga s. (Figure 14.6).

(a) Select Species... and Surface Coverage of ga s in the Contours Of drop-down
list.
(b) Select wall-4 under Surfaces.
(c) Click Display.
Figure 14.6 shows the rate of surface coverage of the site species ga s.

1.63e-01
1.55e-01
1.46e-01
1.38e-01
1.30e-01
1.22e-01
1.14e-01
1.06e-01
9.77e-02
8.95e-02
8.14e-02
7.32e-02
6.51e-02
5.70e-02
4.88e-02
4.07e-02
3.26e-02
2.44e-02
1.63e-02
8.14e-03 Z X
0.00e+00
Y
Contours of Surface Coverage of ga_s

Figure 14.6: Contours of Surface Coverage of ga s
5. Create a line surface from the center of wall-4 to the edge.

Surface −→Line/Rake...
(a) In the Line/Rake Surface panel, select Select Points with Mouse.
(b) In the graphic display, click at the center of wall-4 and at the edge with the
right mouse button.
(c) Click Create.
14-30
6. Plot the surface deposition rate of Ga v/s radial distance(Figure 14.7).

(a) Select Species... and Surface Deposition Rate of ga in the Y Axis Function
drop-down list.
(b) Under Options, deselect Node Values.
The source/sink terms due to the surface reaction are deposited in the cell
adjacent to the wall cells, so it is necessary to plot the cell values and not the
node values.
(c) In the Surfaces list, select line-9.
(d) In the Solution XY Plot, click Plot.
The peak of the surface deposition rate occurs at the center of wall-4 (where the
concentration of the mixture is highest).

line-9
4.00e-05
3.80e-05
3.60e-05
3.40e-05
3.20e-05
Surface
Deposition 3.00e-05
Rate
of 2.80e-05
ga
(kg/m2-s) 2.60e-05
2.40e-05
2.20e-05
-0.02 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14
Z Position (m)
Y
X
Surface Deposition Rate of ga

Figure 14.7: Plot of Surface Deposition Rate of Ga
Extra: You can also perform all the above postprocessing steps to analyze the deposition
of As.
Summary
The main focus of this tutorial is the accurate modeling of macroscopic gas flow, heat
and mass transfer, species diffusion, and chemical reactions (including surface reactions)
in a rotating disk CVD reactor. In this tutorial, you learned how to use the two-step
surface reactions involving site species, and computed simultaneous deposition of gallium
and arsenide from a mixture of precursor gases on a rotating susceptor. Note that the
same approach is valid if you are simulating multi-step reactions with multiple sites/site
species.
14-32
Tutorial 15. Modeling Evaporating Liquid Spray
Introduction
In this tutorial, FLUENT’s air-blast atomizer model is used to predict the behavior of an
evaporating methanol spray. The air flow is modeled first without droplets. To predict
the behavior of the spray, several other discrete-phase models, including collision and
breakup, are used
• Create periodic zones
• Define a spray injection for an air-blast atomizer
• Calculate a solution using FLUENT’s discrete phase model
Prerequisites
Problem Description
The geometry to be considered in this tutorial is shown in Figure 15.1. Methanol is cooled
to −10◦ C before being introduced into an air-blast atomizer. The atomizer contains an
inner air stream surrounded by a swirling annular stream. (The species include the
components of air as well as water vapor, so the model can be expanded to include
combustion, if desired.) To make use of the periodicity of the problem, only a 30-degree
section of the atomizer will be modeled.

Modeling Evaporating Liquid Spray
inner air stream
swirling annular stream
Y
Z
X
Setup and Solution

Preparation
1. Download evaporate_liquid.zip from the Fluent Inc. User Services Center or
2. Unzip evaporate_liquid.zip.
sector.msh can be found in the /evaporate liquid folder created after unzipping
the file.
15-2
Step 1: Grid
1. Read in the mesh file sector.msh.
2. Check the grid.

Grid −→Check

(a) Under Options, select Faces.

(b) Under Surfaces, select only atomizer-wall, central air, and swirling air.
(c) Click the Colors... button.

(d) In the Grid Colors panel, select Color by ID.

This will assign a different color to each zone in the domain, rather than to
each type of zone.
(e) In the Grid Display panel, click Display.
The graphics display will be updated to show the grid. You will now change
the display again to zoom in on an isometric view of the atomizer section.
4. Change the display to an isometric view.

(a) Select right in the Views list and click Restore.

(b) Zoom in and rotate with your mouse to obtain the view shown in Figure 15.2.
15-4
Y
Z
X
Grid
Figure 15.2: Air-Blast Atomizer Mesh Display

5. Using the text interface, change zones periodic-a and periodic-b from wall zones
to periodic zones.
(a) In the console window, type the commands shown in boxes in the dialog below.
> grid
/grid> modify-zones
/grid/modify-zones> list-zones
id name type material kind
---- ---------------- ----------------- ------------------ ----
1 fluid fluid air cell
2 atomizer-wall wall aluminum face
3 central_air mass-flow-inlet face
4 co-flow-air velocity-inlet face
5 outlet pressure-outlet face
6 swirling_air velocity-inlet face
7 periodic-a wall aluminum face
8 periodic-b wall aluminum face
9 outer-wall wall aluminum face
11 default-interior interior face
/grid/modify-zones> make-periodic
Periodic zone [()] 7
Shadow zone [()] 8
Rotational periodic? (if no, translational) [yes] yes
Create periodic zones? [yes] yes
all 1923 faces matched for zones 7 and 8.
zone 8 deleted
created periodic zones.
15-6
6. Reorder the grid twice.

To speed up the solution procedure, the mesh should be reordered, which will sub-
stantially reduce the bandwidth.
Grid −→ Reorder −→Domain
FLUENT will report its progress in the console window:
>> Reordering domain using Reverse Cuthill-McKee method:

zones, cells, faces, done.
Bandwidth reduction = 3286/102 = 32.22
Done.
>> Reordering domain using Reverse Cuthill-McKee method:
zones, cells, faces, done.
Bandwidth reduction = 102/102 = 1
Done.

Step 2: Models

15-8
3. Enable the realizable k- turbulence model.

The realizable k- model gives a more accurate prediction of the spreading rate of
both planar and round jets than the standard k- model.


(a) Select Species Transport under Model.

(b) Choose methyl-alcohol-air in the Mixture Material drop-down list.
The Mixture Material list contains the set of chemical mixtures that exist in
the FLUENT database. By selecting one of the pre-defined mixtures, you are
accessing a complete description of the reacting system. The chemical species
in the system and their physical and thermodynamic properties are defined by
your selection of the mixture material. You can alter the mixture material
selection or modify the mixture material properties using the Materials panel.
When you click OK, the console window will list the properties that are
! required for the models you have enabled. You will see an Information
dialog box, reminding you to confirm the property values that have been
extracted from the database.
(c) Click OK in the Information dialog box to continue.
15-10

1. Set the following conditions for the inner air stream (central air).

2. Set the following conditions for the air stream surrounding the atomizer (co-flow-
air).
15-12
3. Set the following conditions for the exit boundary (outlet).

4. Set the following conditions for the swirling annular stream (swirling air).
15-14
5. Set the following conditions for the outer wall of the atomizer (outer-wall).

Step 4: Initial Solution Without Droplets

The airflow will first be solved and analyzed without droplets.

(a) Select co-flow-air in the Compute From drop-down list.

(b) Click Init to initialize the variables, and then close the panel.
15-16
2. Keep the default under-relaxation factors.




(b) Click OK.
4. Save the case file (spray1.cas).


The solution will converge after about 175 iterations.
6. Save the case and data files (spray1.cas and spray1.dat).

Note: FLUENT will ask you to confirm that the previous case file is to be overwrit-
ten.
15-18
7. Create a clip plane to examine the flow field at the midpoint of the atomizer section.
(a) Select Grid... and Angular Coordinate in the Surface of Constant lists.
(c) Enter 15 in the Iso-Values field.
(d) Enter angle=15 for the New Surface Name.

8. Review the current state of the solution by examining contours of velocity magni-
tude (Figure 15.3).
(a) Select Velocity... and Velocity Magnitude in the Contours Of drop-down list.
(b) Under Options, select Filled and Draw Grid.
15-20
(c) Keep the current grid display settings and close the Grid Display panel.
(d) In the Contours panel, select angle=15 in the Surfaces list.
(e) Click Display.
(f) Use your mouse to obtain the view shown in Figure 15.3.
9.87e+01
9.38e+01
8.88e+01
8.39e+01
7.90e+01
7.40e+01
6.91e+01
6.42e+01
5.92e+01
5.43e+01
4.94e+01
4.44e+01
3.95e+01
3.45e+01
2.96e+01
2.47e+01
1.97e+01
1.48e+01
9.87e+00 Y
4.94e+00 Z
0.00e+00 X
Contours of Velocity Magnitude (m/s)

FLUENT 6.2 (3d, segregated, spe, rke)
Figure 15.3: Velocity Magnitude at Mid-Point of Atomizer Section

9. Display path lines of the air in the swirling annular stream (Figure 15.4).
Display −→Path Lines...
(a) In the Release From Surfaces list, select swirling air.

You will need to scroll down to access this item.
(b) Increase the Path Skip value to 5.
(c) Under Options, select Draw Grid.
(d) Keep the current grid display settings and close the Grid Display panel.
(e) Click Display in the Path Lines panel.
(f) Use your mouse to obtain the view shown in Figure 15.4.
15-22
4.90e+01
4.65e+01
4.41e+01
4.17e+01
3.92e+01
3.68e+01
3.43e+01
3.19e+01
2.94e+01
2.70e+01
2.45e+01
2.20e+01
1.96e+01
1.71e+01
1.47e+01
1.22e+01
9.80e+00
7.35e+00
4.90e+00 Y
2.45e+00 Z
0.00e+00 X
Path Lines Colored by Particle ID

Figure 15.4: Path Lines of Air in the Swirling Annular Stream

Step 5: Create a Spray Injection

1. Define the discrete phase modeling parameters.
Define −→ Models −→Discrete Phase...
(a) Define the interphase interaction.

i. Under Interaction, turn on Interaction with Continuous Phase.
This will include the effects of the discrete phase trajectories on the con-
tinuous phase.
ii. Specify a value of 20 for Number of Continuous Phase Iterations per DPM
Iteration.
(b) Set the Physical Models.
i. Select the Physical Models tab.
ii. Under Spray Model, select Droplet Collision and Droplet Breakup.
15-24
iii. Under Breakup Model, keep the default selection of TAB.

iv. Under Breakup Constants, keep the default value of 0 for y0.
This parameter is the dimensionless droplet distortion at t = 0.
(c) Specify the Tracking Parameters.
i. Select the Tracking tab.

ii. Keep the default value of Step Length Factor.
iii. Under Drag Parameters, select dynamic-drag in the Drag Law drop-down
list.
The dynamic-drag law is available only when the Droplet Breakup model is
used.

(d) Set the particle treatment.

i. Under Particle Treatment, select Unsteady Particle Tracking.
ii. Specify a value of 0.001 for Particle Time Step Size.
iii. Retain the default value of 1 for Number of Time Steps.
2. Create the spray injection.

In this step, you will define the characteristics of the atomizer.
Define −→Injections...
(a) Click the Create button at the top of the panel.

This will open the Set Injection Properties panel.
15-26
(b) In the Injection Type drop-down list, select air-blast-atomizer.

(c) Increase the Number Of Particle Streams to 60.
This option controls how many droplet parcels are introduced into the domain
at every time step.
(d) Under Particle Type, select Droplet.
(e) In the Material drop-down list, select methyl-alcohol-liquid.
(f) Set the point properties for the injection.
Use scroll down bar to see all the point properties.
i. Set the X-Position, Y-Position, and Z-Position of the injection to 0, 0, and
0.0015.
ii. Set the X-Axis, Y-Axis, and Z-Axis of the injection to 0, 0, and 1.
iii. Set the Temperature to 263 K.

iv. Set the Flow Rate to 1.7e-4 kg/s.

This is the methanol flow rate for a 30-degree section of the atomizer. The
actual atomizer flow rate is 12 times this value.
v. Keep the default Start Time of 0 s and set the Stop Time to 100 s.
For this problem, the injection should begin at t = 0 and not stop until
long after the time period of interest. A large value for the stop time (e.g.,
100 s) will ensure that the injection will essentially never stop.
vi. Set the Injector Inner Diam. to 0.0035 m, and the Injector Outer Diam. to
0.0045 m.
vii. Set the Spray Half Angle to -45 deg.
The spray angle is the angle between the liquid sheet trajectory and the
injector centerline. In this case, the value is negative because the sheet is
initially converging toward the centerline.
viii. Set the Relative Velocity to 82.6 m/s.
The relative velocity is the expected relative velocity between the atomizing
air and the liquid sheet.
ix. Keep the default Azimuthal Start Angle of 0 deg and set the Azimuthal
Stop Angle to 30 deg.
This will restrict the injection to the 30-degree section of the atomizer that
is being modeled.
(g) Define the turbulent dispersion.
i. Click the Turbulent Dispersion tab.
The lower half of the panel will change to show options for the turbulent
dispersion model.
ii. Under Stochastic Tracking, turn on the Discrete Random Walk Model and
Random Eddy Lifetime options.
These models will account for the turbulent dispersion of the droplets.
Note: In the case that the spray injection would be striking a wall, you would need
to specify the wall boundary conditions for the droplets. Though this tutorial
does have wall zones, they are a part of the atomizer apparatus. Because these
walls are not in the path of the spray droplets, you do not need to change the
wall boundary conditions any further.
15-28
3. Set the droplet material properties.

Because the secondary atomization models (breakup and coalescence) are used, the
droplet properties must be set.
(a) In the Material Type drop-down list, select droplet-particle.

(b) Under Properties, enter a value of 0.0056 kg/m-s for Viscosity.
(c) Under Properties, scroll down and change Saturation Vapor Pressure to
piecewise-linear. Keep the default values.
(d) Under Properties, scroll down and enter a value of 0.0222 N/m for Droplet
Surface Tension.
(e) Click Change/Create to accept the change in properties for the methanol
droplet material.

Step 6: Solution
1. Uncheck the convergence criteria for all the residuals.
Solve −→ Monitors −→Residuals...
2. Reduce the Under-Relaxation Factor for Discrete Phase Sources to 0.1.
3. Request for 200 iterations.
4. Save the case and data files (spray2.cas and spray2.dat).
5. Display the trajectories of the droplets in the spray injection (Figure 15.5).
This will allow you to review the location of the droplets.
(a) In the Style drop-down list, deep the default of point.

(b) Click the Style Attributes... button.
This will open the Path Style Attributes panel.
15-30
(c) Set the Marker Size to 0.25 and click OK.

(d) In the Particle Tracks panel, select Draw Grid under Options.
(e) Keep the current display settings and close the panel.
(f) In the Particle Tracks panel, select Particle Variables... and Particle Diameter in
the Color by drop-down list.
This will display the location of the droplets colored by their diameters.
(g) In the Release from Injections list, select injection-0.
(h) Click Display.
(i) Use your mouse to obtain the view shown in Figure 15.5.

1.43e-04
1.37e-04
1.31e-04
1.24e-04
1.18e-04
1.12e-04
1.06e-04
9.93e-05
9.31e-05
8.68e-05
8.05e-05
7.42e-05
6.79e-05
6.17e-05
5.54e-05
4.91e-05
4.28e-05
3.65e-05
3.03e-05
Z
2.40e-05 Y
1.77e-05 X
Particle Traces Colored by Particle Diameter (m)

Figure 15.5: Particle Tracks for the Spray Injection After 200 Iterations
The air-blast atomizer model assumes that a cylindrical liquid sheet exits the
atomizer, which then disintegrates into ligaments and droplets. Appropriately,
the model determines that the droplets should be input into the domain in a
ring. The radius of this disk is determined from the inner and outer radii of
the injector.
Note that the maximum diameter of the droplets is about
10−4 m, or 0.1 mm. This is slightly smaller than the film height, which makes
sense. Recall that the inner diameter and outer diameter of the injector are
3.5 mm and 4.5 mm, respectively. The film height is then 12 (4.5 − 3.5) =
0.5 mm. The range in the droplet sizes is due to the fact that the air-blast
atomizer automatically uses a distribution of droplet sizes.
Also note that the droplets are placed a slight distance away from the injector.
Once the droplets are injected into the domain, they can collide/coalesce with
other droplets as determined by the secondary models (breakup and collision).
However, once a droplet has been introduced into the domain, the air-blast
atomizer model no longer affects the droplet.

7. Save the new case and data files (spray3.cas and spray3.dat).
15-32
1. Display the particle trajectories again, to see how the droplets have dispersed.
(a) Click Display in the Particle Tracks panel.
(b) Use your mouse to obtain the view shown in Figure 15.6.
1.76e-04
1.68e-04
1.60e-04
1.52e-04
1.44e-04
1.36e-04
1.28e-04
1.20e-04
1.13e-04
1.05e-04
9.67e-05
8.88e-05
8.09e-05
7.30e-05
6.50e-05
5.71e-05
4.92e-05
4.13e-05
3.34e-05 Y
Z
2.54e-05
1.75e-05 X
Particle Traces Colored by Particle Diameter (m)

Figure 15.6: Particle Tracks for the Spray Injection After 400 Iterations.
2. Create an isosurface of the methanol mass fraction.


(a) Select Species... and Mass fraction of ch3oh in the Surface of Constant lists.
(c) Enter 0.0015 in the Iso-Values field.
(d) Enter methanol-mf=0.0015 for the New Surface Name.
3. Display the isosurface you just created (methanol-mf=0.0015).

15-34
(a) Select methanol-mf=0.0015 in the Surfaces list.

(b) Click the Colors... button.
(c) In the Grid Colors panel, select Color By Type.

(d) Scroll down and select surface in the Types list and dark red in the Colors list.
This will ensure that the isosurface is displayed in red, which contrasts better
with the rest of the grid.
(e) In the Grid Display panel, click Display.
The graphics display will be updated to show the isosurface.
4. Modify the view to include the entire atomizer.

(a) Click Define... to open the Graphics Periodicity panel.

i. Under Cell Zones, select fluid.

ii. Make sure Rotational is selected under Periodic Type.
iii. Increase the Number of Repeats to 12.
iv. Click Set and close the Graphics Periodicity panel.
v. Click Apply in the Views panel.
vi. In the Grid Display panel, click Display.
The graphics display will be updated to show the entire atomizer.
vii. Use your mouse to obtain the view shown in Figure 15.7.
15-36
XY
Z
Grid
Figure 15.7: Full Atomizer Display with Surface of Constant Methanol Mass Fraction
Summary
In this tutorial, you defined a spray injection for an air-blast atomizer and calculated
the solution using FLUENT’s under relaxation factor for discrete phase source. You
viewed the location of methanol droplet particles after they had exited the atomizer and
examined an isosurface of the methanol mass fraction.

15-38
Tutorial 16. Using the VOF Model
Introduction
This tutorial illustrates the setup and solution of the two-dimensional turbulent fluid
flow in a partially filled spinning bowl.
• Set up and solve a transient free-surface problem using the segregated solver
• Model the effect of gravity
• Copy a material from the property database
• Patch initial conditions in a subset of the domain
• Define a custom field function
• Mirror and rotate the view in the graphics window
• Examine the fluid flow and the free-surface shape using velocity vectors and volume
fraction contours
Prerequisites
This tutorial requires a basic familiarity with FLUENT. You may also find it helpful to
read about VOF multiphase flow modeling in the FLUENT by reading Section 24.2 of the
User’s Guide for more information. Otherwise, no previous experience with multiphase
modeling is required.
Problem Description
The information relevant to this problem is shown in Figure 16.1. A large bowl, 1 m in
radius, is one-third filled with water and is open to the atmosphere. The bowl spins with
an angular velocity of 3 rad/sec. Based on the rotating water, the Reynolds number is
about 106 , so the flow is modeled as turbulent.

Using the VOF Model
2m
1m
Bowl: Ω = 3 rad/s
Air: ρ = 1.225 kg/m 3
-5
µ = 1.7894 x 10 kg/m-s
Water: ρ = 998.2 kg/m 3
µ = 1 x 10
-3
kg/m-s
Figure 16.1: Water and Air in a Spinning Bowl
Setup and Solution

Preparation
1. Download vof.zip from the Fluent Inc. User Services Center or copy it from the
FLUENT documentation CD to your working directory (as described in Tutorial 1).
2. Unzip vof.zip.
bowl.msh can be found in the /vof folder created after unzipping the file.
The mesh file bowl.msh is a quadrilateral mesh describing the system geometry
shown in Figure 16.1.
16-2
Using the VOF Model
Step 1: Grid
1. Read the 2D grid file, bowl.msh.

As shown in Figure 16.2, half of the bowl is modeled, with a symmetry boundary at
the centerline. The bowl is shown lying on its side, with the region to be modeled
extending from the centerline to the outer wall. When you begin to display data
graphically, you will need to rotate the view and mirror it across the centerline to
obtain a more realistic view of the model. This step will be performed later in the
tutorial.

Using the VOF Model
Grid
16-4
Using the VOF Model
Step 2: Models
1. Specify a transient model with axisymmetric swirl.

The segregated solver must be used for multiphase calculations.
(b) Under Space, select Axisymmetric Swirl.
(c) Under Time, select Unsteady.

Using the VOF Model
2. Turn on the VOF model.

Define −→ Models −→Multiphase...
(a) Select Volume of Fluid as the Model.

The panel will expand to show inputs for the VOF model.
(b) Under VOF Parameters, select Geo-Reconstruct (the default) as the VOF Scheme.
This is the most accurate interface-tracking scheme, and is recommended for
most transient VOF calculations.
When you click OK, FLUENT will report that one of the zone types will need
to be changed before proceeding with the calculation. You will take care of this
step when you input boundary conditions for the problem.
16-6
Using the VOF Model

(a) Select k-epsilon as the Model, and retain the default setting of Standard under
k-epsilon Model.

Using the VOF Model
Step 3: Materials
1. Copy water from the FLUENT database materials so that it can be used for the
secondary phase.
(a) Click on the Fluent Database... button to open the Fluent Database Materials
panel.
(b) In the Fluent Fluid Materials list (near the bottom), select water-liquid.
(c) Click Copy and close the Fluent Database Materials and Materials panels.
16-8
Using the VOF Model
Step 4: Phases
Here, water is defined as the secondary phase mainly for convenience in setting up the
problem. When you define the initial solution, you will be patching an initial swirl velocity
in the bottom third of the bowl, where the water is. It is more convenient to patch a water
volume fraction of 1 there than to patch an air volume fraction of 1 in the rest of the
domain. Also, the default volume fraction at the pressure inlet is 0, which is the correct
value if water is the secondary phase.
In general, you can specify the primary and secondary phases whichever way you prefer.
It is a good idea, especially in more complicated problems, to consider how your choice
will affect the ease of problem setup.
1. Define the air and water phases within the bowl.

Define −→Phases...
(a) Specify air as the primary phase.

i. Select phase-1 and click the Set... button.
ii. In the Primary Phase panel, enter air for the Name.
iii. Keep the default selection of air for the Phase Material.

Using the VOF Model
(b) Specify water as the secondary phase.

ii. In the Secondary Phase panel, enter water for the Name.
iii. Select water-liquid from the Phase Material drop-down list.
16-10
Using the VOF Model

1. Set the gravitational acceleration.

(b) Set the Gravitational Acceleration in the X direction to 9.81 m/s2 .
Since the centerline of the bowl is the x axis, gravity points in the positive x
direction.
2. Set the operating density.

(a) Under Variable-Density Parameters, turn on the Specified Operating Density op-
tion and accept the Operating Density of 1.225.
It is a good idea to set the operating density to be the density of the lighter
phase. This excludes the buildup of hydrostatic pressure within the lighter
phase, improving the round-off accuracy for the momentum balance.
Note: The Reference Pressure Location (0,0) is situated in a region where the fluid
will always be 100% of one of the phases (air), a condition that is essential
for smooth and rapid convergence. If it were not, you would need to change it
to a more appropriate location.

Using the VOF Model

1. Change the bowl centerline from a symmetry boundary to an axis boundary.

For axisymmetric models, the axis of symmetry must be an axis zone.
(a) Select symmetry-2 in the Zone list in the Boundary Conditions panel.
(b) In the Type list, choose axis.
You will have to scroll to the top of the list.
(c) Click Yes in the Question dialog box that appears.
(d) Click OK in the Axis panel to accept the default Zone Name.
2. Set the conditions at the top of the bowl (pressure-inlet-4).

For the VOF model, you will specify conditions for the mixture (i.e., conditions that
apply to all phases) and also conditions that are specific to the secondary phase.
There are no conditions to be specified for the primary phase.
(a) Set the conditions for the mixture.
i. In the Boundary Conditions panel, keep the default selection of mixture in
the Phase drop-down list and click Set....
16-12
Using the VOF Model
ii. Set the Turb. Kinetic Energy to 2.25e-2 and the Turb. Dissipation Rate to
7.92e-3.
Since there is initially no flow passing through the pressure inlet, you need
to specify k and explicitly rather than using one of the other turbulence
specification methods. All of the other methods require you to specify the
turbulence intensity, which is 0 in this case.
The values for k and are computed as follows:
k = (Iwwall )2
0.093/4 k 3/2
=
`
where the turbulence intensity I is 0.05 (close to zero), wwall is 3 m/s,
and ` is 0.07 (obtained by multiplying 0.07 by the maximum radius of the
bowl, which is 1).
See Section 7.2.2 of the User’s Guide for details about the specification of
turbulence boundary conditions at flow inlets and exits.
(b) Check the volume fraction of the secondary phase.
i. In the Boundary Conditions panel, select water from the Phase drop-down
list and click Set....

Using the VOF Model
ii. Retain the default Volume Fraction of 0.

A water volume fraction of 0 indicates that only air is present at the
pressure inlet.
3. Set the conditions for the spinning bowl (wall-1).

For a wall boundary, all conditions are specified for the mixture. There are no
conditions to be specified for the individual phases.
(a) In the Boundary Conditions panel, select mixture in the Phase drop-down list
and click Set....
16-14
Using the VOF Model
(b) Select Moving Wall under Wall Motion.

The panel will expand to show inputs for the wall motion.
(c) Under Motion, choose Rotational and then set the rotational Speed (rad/s) to
3.

Using the VOF Model
Step 7: Solution
In simple flows, the under-relaxation factors can usually be increased at the start of the
calculation. This is particularly true when the VOF model is used, where high under-
relaxation on all variables can greatly improve the performance of the solver.

(a) Set all Under-Relaxation Factors to 1.
Be sure to use the scroll bar to access the under-relaxation factors that are
! initially out of view.
(b) Under Discretization, choose the PRESTO! scheme in the drop-down list next
to Pressure.
(c) Under Pressure-Velocity Coupling, select PISO.
PISO is recommended for transient flow calculations.
16-16
Using the VOF Model
2. Enable the display of residuals during the solution process.


(b) Click OK button to close the panel.
3. Enable the plotting of the axial velocity of water near the outer edge of the bowl
during the calculation.
For transient calculations, it is often useful to monitor the value of a particular
variable to see how it changes over time. Here you will first specify the point at
which you want to track the velocity, and then define the monitoring parameters.
(a) Define a point surface near the outer edge of the bowl.

Using the VOF Model
i. Set the x0 and y0 coordinates to 0.75 and 0.65.

ii. Enter point for the New Surface Name.
iii. Click Create.
(b) Define the monitoring parameters.
i. Increase the Surface Monitors value to 1.

ii. Turn on the Plot and Write options for monitor-1.
velocity history will be written to a file. If you do not select the Write
option, the history information will be lost when you exit FLUENT.
iii. In the drop-down list under Every, choose Time Step.
16-18
Using the VOF Model
iv. Click on Define... to specify the surface monitor parameters in the Define
v. Select Vertex Average from the Report Type drop-down list.

This is the recommended choice when you are monitoring the value at a
single point using a point surface.
vi. Select Flow Time in the X Axis drop-down list.
vii. Select Velocity... and Axial Velocity in the Report Of drop-down lists.
viii. Select point in the Surfaces list.
ix. Enter axial-velocity.out for the File Name.
x. Click OK in the Define Surface Monitor panel and then in the Surface
Monitors panel.

Using the VOF Model

(a) Select pressure-inlet-4 in the Compute From drop-down list.

All initial values will be set to zero, except for the turbulence quantities.
(b) Click Init and close the panel.
5. Patch the initial distribution of water (i.e., water volume fraction of 1.0) and a
swirl velocity of 3 rad/s in the bottom third of the bowl (where the water is).
In order to patch a value in just a portion of the domain, you will need to define
a cell “register” for that region. You will use the same tool that is used to mark a
region of cells for adaption. Also, you will need to define a custom function for the
swirl velocity.
(a) Define a register for the bottom third of the domain.
Adapt −→Region...
16-20
Using the VOF Model
i. Set the (Xminimum,Yminimum) coordinate to (0.66,0), and the (Xmaxi-

mum,Ymaximum) coordinate to (1,1).
ii. Click the Mark button.
This creates a register containing the cells in this region.
(b) Check the register to be sure it is correct.
Adapt −→Manage...
i. Select the register (hexahedron-r0) in the Registers list and click Display.
The graphics display will show the bottom third of the bowl in red.

Using the VOF Model
6. Define a custom field function for the swirl velocity w = 3r.

(a) Click the 3 button on the calculator pad.

The 3 will appear in the Definition field. If you make a mistake, click the DEL
button to delete the last item you added to the function definition.
(b) Click the X button on the calculator pad.
(c) In the Field Functions drop-down list, select Grid... and Radial Coordinate.
(d) Click the Select button.
radial-coordinate will appear in the Definition.
(e) Enter a New Function Name of swirl-init.
(f) Click Define.
Note: If you wish to check the function definition, click on the Manage...
button and select swirl-init.
16-22
Using the VOF Model
7. Patch the water volume fraction in the bottom third of the bowl.
Solve −→ Initialize −→Patch...
(a) In the Phase drop-down list, select water.

(b) Select Volume Fraction in the Variable list.
(c) Select hexahedron-r0 in the Registers to Patch list.
(d) Set the Value to 1.
(e) Click Patch.
This sets the water volume fraction to 1 in the lower third of the bowl. That
is, you have defined the lower third of the bowl to be filled with water.

Using the VOF Model
(f) Patch the swirl velocity in the bottom third of the bowl.
i. In the Phase drop-down list, select mixture.

ii. Choose Swirl Velocity in the Variable list.
iii. Enable the Use Field Function option and select swirl-init in the Field Func-
tion list.
iv. Click Patch.
It’s a good idea to check your patch by displaying contours of the patched fields.
16-24
Using the VOF Model
8. Display contours of swirl velocity.

(a) Select Velocity... and Swirl Velocity in the Contours of lists.

(b) Enable the Filled option and turn off the Node Values option.
Since the values you patched are cell values, you should view the cell values,
rather than the node values, to check that the patch has been performed cor-
rectly. (FLUENT computes the node values by averaging the cell values.)
(c) Click Display.
To make the view more realistic, you will need to rotate the display and mirror it
across the centerline.

Using the VOF Model
9. Rotate the view and mirror it across the centerline.

(a) Select axis-2 in the Mirror Planes list and click Apply.
(b) Use your middle and left mouse buttons to zoom and translate the view so
that the entire bowl is visible in the graphics display.
(c) Click on the Camera... button to open the Camera Parameters panel.
(d) Using your left mouse button, rotate the dial clockwise until the bowl appears
upright in the graphics window (90◦ ).
(e) Close the Camera Parameters panel.
(f) In the Views panel, click on the Save button under Actions to save the mirrored,
upright view, and then close the panel.
When you do this, view-0 will be added to the list of Views.
The upright view of the bowl in Figure 16.3 correctly shows that w = 3r in the
region of the bowl that is filled with water.
16-26
Using the VOF Model
2.35e+00
2.23e+00
2.12e+00
2.00e+00
1.88e+00
1.76e+00
1.65e+00
1.53e+00
1.41e+00
1.29e+00
1.18e+00
1.06e+00
9.41e-01
8.23e-01
7.06e-01
5.88e-01
4.70e-01
3.53e-01
2.35e-01
1.18e-01
0.00e+00
Contours of Swirl Velocity (mixture) (m/s) (Time=0.0000e+00)

FLUENT 6.2 (axi, swirl, segregated, vof, ske, unsteady)
Figure 16.3: Contours of Initial Swirl Velocity

Using the VOF Model
10. Display contours of water volume fraction.
(a) Select Phases... and Volume fraction of water in the Contours of lists.
(b) Select water in the Phase drop-down list.
(c) Set the number of contour Levels to 2 and click Display.
There are only two possible values for the volume fraction at this point: 0 or
1.
16-28
Using the VOF Model
1.00e+00
5.00e-01
0.00e+00
Contours of Volume fraction (water) (Time=0.0000e+00)

Figure 16.4: Contours of Initial Water Volume Fraction
Figure 16.4 correctly shows that the bottom third of the bowl contains water.
11. Set the time-step parameters for the calculation.

(a) Under Time, specify a value of 0.002 for Time Step Size and a value of 1000
for Number of Time Steps.
(b) Under Time Stepping Method, select Variable.
(c) Under Variable Time Step Parameters, specify a value of 0.002 for Minimum
Time Step Size and a value of 0.01 for Maximum Time Step Size.
(d) Retain the other default parameters.
(e) Click Apply.
This will save the time step size to the case file (the next time a case file is
saved).
(f) Save the initial case and data files (bowl.cas and bowl.dat).
(g) Specify a value of 0.4 for Ending Time.
As iterations will begin with variable time step method, a value of 0.4 for
Ending Time will stop the calculations after t = 0.4sec. Save the data file at
this moment and proceed the calculations for Ending Time of 0.6, 0.8, 1.0, and
2.0. You may have to reset the value of Time Step Size to a value of 0.002
after saving each data file.

Using the VOF Model
Figure 16.5 shows the time history for the axial velocity. The velocity is clearly
oscillating, and the oscillations appear to be decaying over time (as the peaks
become smaller). This periodic oscillation has a cycle of 1 second. The switch
from a positive to a negative axial velocity indicates that the water is sloshing
up and down the sides of the bowl in an attempt to reach an equilibrium po-
sition. The fact that the amplitude is decaying suggests that equilibrium will
be reached at some point. The periodic behavior in evidence will therefore be
present only during the initial startup phase of the bowl rotation.
0.3000
0.2000
0.1000
Vertex 0.0000
Average
Axial
Velocity -0.1000
-0.2000
-0.3000
0.0000 0.2000 0.4000 0.6000 0.8000 1.0000 1.2000 1.4000 1.6000 1.8000 2.0000
Flow Time
Convergence history of Axial Velocity on point (in SI units) (Time=2.0000e+00)

Figure 16.5: Time History of Axial Velocity
16-30
Using the VOF Model
As indicated by changes in axial velocity in Figure 16.5, the flow field is oscillating peri-
odically. In this step, you will examine the flow field at several different times. (Recall
that you have saved the data files for t = 0.4, 0.6, 0.8, and, 1.0.)
1. Read in the data file of interest.

File −→ Read −→Data...
2. Display filled contours of water volume fraction.

Hint: Follow the instructions in substep 5h of Step 7: Solution (on page 16-28),
but turn Node Values back on.
Figures 16.6–16.9 show that the water level decreases from t = 0.4 to t = 0.6, then
increases from t = 0.6 to t = 1. At t = 1, the water level in the center of the
bowl has risen above the initial level, so you can expect the cycle to repeat as the
water level begins to decrease again in an attempt to return to equilibrium. (You
can read in the data files between t = 1 and t = 2 to confirm that this is in fact
what happens.
Since the time history of axial velocity (Figure 16.5) shows that the velocity os-
cillation is decaying over time, you can expect that if you were to continue the
calculation, the water level would eventually reach some point where the gravita-
tional and centrifugal forces balance and the water level reaches a new equilibrium
point.
Extra: Try continuing the calculation to determine how long it takes for the axial
velocity oscillations in Figure 16.5 to disappear.

Using the VOF Model
1.00e+00
5.00e-01
0.00e+00
Contours of Volume fraction (water) (Time=4.0000e-01)

Figure 16.6: Shape of the Free Surface at t = 0.4
1.00e+00
5.00e-01
0.00e+00

16-32
Using the VOF Model
1.00e+00
5.00e-01
0.00e+00

1.00e+00
5.00e-01
0.00e+00
Contours of Volume fraction (water) (Time=1.0000e+00)

Figure 16.9: Shape of the Free Surface at t = 1

Using the VOF Model
3. Plot contours of stream function.

(a) Select Stream Function (in the Velocity... category) in the Contours of drop-
down list.
(b) Turn off the Filled option and increase the number of contour Levels to 30.
(c) Click on Display.
In Figures 16.10–16.13, you can see a recirculation region that falls and rises as
the water level changes. To get a better sense of these recirculating patterns, you
will next look at velocity vectors.
16-34
Using the VOF Model
2.77e+01
2.59e+01
2.40e+01
2.22e+01
2.03e+01
1.85e+01
1.66e+01
1.48e+01
1.29e+01
1.11e+01
9.25e+00
7.40e+00
5.55e+00
3.70e+00
1.85e+00
0.00e+00
Contours of Stream Function (mixture) (kg/s) (Time=4.0000e-01)

Figure 16.10: Contours of Stream Function at t = 0.4
2.47e+01
2.30e+01
2.14e+01
1.97e+01
1.81e+01
1.64e+01
1.48e+01
1.31e+01
1.15e+01
9.86e+00
8.22e+00
6.57e+00
4.93e+00
3.29e+00
1.64e+00
0.00e+00


Using the VOF Model
4.68e+01
4.37e+01
4.06e+01
3.74e+01
3.43e+01
3.12e+01
2.81e+01
2.50e+01
2.18e+01
1.87e+01
1.56e+01
1.25e+01
9.36e+00
6.24e+00
3.12e+00
0.00e+00

7.02e+00
6.55e+00
6.08e+00
5.61e+00
5.14e+00
4.68e+00
4.21e+00
3.74e+00
3.27e+00
2.81e+00
2.34e+00
1.87e+00
1.40e+00
9.35e-01
4.68e-01
0.00e+00
Contours of Stream Function (mixture) (kg/s) (Time=1.0000e+00)

Figure 16.13: Contours of Stream Function at t = 1
16-36
Using the VOF Model
4. Plot velocity vectors in the bowl.

(a) In the Style drop-down list, select arrow.

This will make the velocity direction easier to see.
(b) Increase the Scale factor to 6 and increase the Skip value to 1.
(c) Click on Vector Options... to open the Vector Options panel.

Using the VOF Model
i. Turn off the Z Component.

This allows you to examine the non-swirling components only.
ii. Click Apply and close the panel.
(d) Click on Display.
Figures 16.14–16.17 show the changes in water and air flow patterns between t = 0.4
and t = 1. In Figure 16.14, you can see that the flow in the middle of the bowl is
being pulled down by gravitational forces, and pushed out and up along the sides of
the bowl by centrifugal forces. This causes the water level to decrease in the center
of the bowl, as shown in the volume fraction contour plots, and also results in the
formation of a recirculation region in the air above the water surface.
In Figure 16.15, the flow has reversed direction, and is slowly rising up in the mid-
dle of the bowl and being pulled down along the sides of the bowl. This reversal
occurs because the earlier flow pattern caused the water to overshoot the equilib-
rium position. The gravity and centrifugal forces now act to compensate for this
overshoot.
16-38
Using the VOF Model
1.91e+00
1.82e+00
1.72e+00
1.63e+00
1.53e+00
1.44e+00
1.34e+00
1.25e+00
1.15e+00
1.06e+00
9.61e-01
8.66e-01
7.71e-01
6.75e-01
5.80e-01
4.85e-01
3.89e-01
2.94e-01
1.99e-01
1.03e-01
7.98e-03
Velocity Vectors Colored By Velocity Magnitude (mixture) (m/s) (Time=4.0000e-01)

Figure 16.14: Velocity Vectors for the Air and Water at t = 0.4
1.96e+00
1.87e+00
1.77e+00
1.67e+00
1.57e+00
1.47e+00
1.37e+00
1.28e+00
1.18e+00
1.08e+00
9.82e-01
8.84e-01
7.86e-01
6.88e-01
5.90e-01
4.91e-01
3.93e-01
2.95e-01
1.97e-01
9.88e-02
6.55e-04


Using the VOF Model
2.12e+00
2.01e+00
1.91e+00
1.80e+00
1.70e+00
1.59e+00
1.48e+00
1.38e+00
1.27e+00
1.17e+00
1.06e+00
9.56e-01
8.51e-01
7.45e-01
6.39e-01
5.34e-01
4.28e-01
3.22e-01
2.17e-01
1.11e-01
5.18e-03

2.12e+00
2.01e+00
1.91e+00
1.80e+00
1.70e+00
1.59e+00
1.48e+00
1.38e+00
1.27e+00
1.17e+00
1.06e+00
9.55e-01
8.49e-01
7.44e-01
6.38e-01
5.32e-01
4.26e-01
3.20e-01
2.14e-01
1.08e-01
2.22e-03
Velocity Vectors Colored By Velocity Magnitude (mixture) (m/s) (Time=1.0000e+00)

Figure 16.17: Velocity Vectors for the Air and Water at t = 1
16-40
Using the VOF Model
In Figure 16.16 you can see that the flow is rising up more quickly in the middle of
the bowl, and in Figure 16.17 you can see that the flow is still moving upward, but
more slowly. These patterns correspond to the volume fraction plots at these times.
As the upward motion in the center of the bowl decreases, you can expect the flow
to reverse as the water again seeks to reach a state of equilibrium.
Summary
In this tutorial, you have learned how to use the VOF free surface model to solve a
problem involving a spinning bowl of water. The time-dependent VOF formulation is
used in this problem to track the shape of the free surface and the flow field inside the
spinning bowl.
You observed the changing pattern of the water and air in the bowl by displaying volume
fraction contours, stream function contours, and velocity vectors at t = 0.4, t = 0.6,
t = 0.8, and t = 1 second.

Using the VOF Model
16-42
Tutorial 17. Modeling Cavitation
Introduction
This tutorial examines the pressure-driven cavitating flow of water
through a sharp-edged orifice. This is a typical configuration in fuel injectors, and brings
a challenge to the physics and numerics of cavitation models, because of the high pres-
sure differentials involved, and the high ratio of liquid to vapor density. Using FLUENT’s
multiphase modeling capability, you will be able to predict the strong cavitation near the
orifice after flow separation at a sharp edge. In this tutorial you will learn how to:
• Set boundary conditions for internal flow
• Use the mixture model with cavitation effects
Prerequisites
Problem Description
The problem considers the cavitation caused by the flow separation after a sharp-edged
orifice. The flow is pressure driven, with an inlet pressure of 5 × 105 Pa, and an outlet
pressure of 9.5 × 104 Pa. The orifice diameter is 4 × 10− 3 m, and geometrical param-
eters of the orifice are D/d = 2.88 and L/r = 8, where D, d, and L are inlet diameter,
orifice diameter, and orifice length respectively. The geometry of the orifice is shown in
Figure 17.1.

Modeling Cavitation
Setup and Solution

Preparation
1. Download cavitation.zip from the Fluent Inc. User Services Center or copy it
from the FLUENT documentation CD to your working directory (as described in
Tutorial 1).
2. Unzip cavitation.zip.
cav.msh can be found in the /cavitation folder created after unzipping the file.
17-2
Modeling Cavitation
Step 1: Grid
1. Read the grid file (cav.msh).
2. Check the grid.

Grid −→Check
3. Check the grid scale.

Grid −→Scale...
(a) Keep the default settings and close the panel.

(a) Display the grid using the default settings (Figure 17.2).
As shown in Figure 17.2, half of the problem geometry is modeled, with an
axis boundary (consisting of two separate lines) at the centerline. Especially
when you begin to display data graphically, you may want to mirror the view
across the centerline to obtain a more realistic view of the model. This step
will be performed later in the tutorial.
The mesh is quadrilateral, slightly graded in the plenum to be finer toward the
orifice. In the orifice, the mesh is uniform, with aspect ratios close to 1, as
the flow is expected to exhibit two-dimensional gradients.

Modeling Cavitation
Grid
Figure 17.2: The Grid in the Orifice
17-4
Modeling Cavitation
Step 2: Models
1. Specify a steady-state axisymmetric model.
(a) Under Space, select Axisymmetric.

(b) Keep the default settings for everything else.
Note: A computationally-intensive unsteady calculation is necessary to accurately
simulate the irregular cyclic process of bubble formation, growth, filling by
water jet re-entry, and breakoff. In this tutorial, you will perform a steady-
state calculation to simulate the presence of a bubble in the separation region
in the time-averaged flow.

Modeling Cavitation
2. Enable the multiphase mixture model.

(a) Select Mixture as the Model.

(b) Under Mixture Parameters, turn off the Slip Velocity option.
Since there is no significant difference in velocities for the different phases,
there is no need to solve for the slip velocity equation.
3. Enable cavitation effect.
(a) Click Interaction...

This will open the Phase Interaction panel.
17-6
Modeling Cavitation
i. Activate Cavitation.
The panel will expand again to show the cavitation inputs.
ii. Enter 3540 for the Vaporization Pressure (pascal).
The vaporization pressure is a property of the working liquid, which de-
pends mainly on the liquid temperature. The default value is the vapor-
ization pressure of water at a temperature of 300 K.
iii. Enter 0.0717 for the Surface Tension Coefficient.
Like the vaporization pressure, the liquid-vapor surface tension is a prop-
erty of the liquid, which depends mainly on temperature. Here too, the
default value is the surface tension for water and vapor at a temperature
of 300 K.
iv. Enter 1.5e-5 for Non-Condensable Gas Mass Fraction .
This is the mass fraction of non condensable gas dissolved in the working
liquid. 1.5e − 5 (15 ppm) is a typical value for air dissolved in water.
(a) Select k-epsilon as the Model.

Modeling Cavitation
(b) Keep the default selection of Standard under k-epsilon Model.

The standard k-model used in conjunction with standard wall functions is a
suitable choice for this type of mesh resolution.
For different cavitation problems, you may use other turbulence models.
See Section 24.6.4 of the User’s Guide for more information on the choice of
turbulence models to be used in conjunction with FLUENT’s cavitation model.
(c) Keep the default selection of Standard Wall Functions under Near-Wall Treat-
ment.
17-8
Modeling Cavitation
Step 3: Materials
1. Create a new material to be used for the primary phase. Copy water vapor from
the materials database so that it can be used for the secondary phase, and modify
its density.
(a) In the Name field, type water.
(b) Clear the Chemical Formula field.
(c) In the Density drop-down list, keep the default selection of constant, and enter
a value of 1000.
(d) In the Viscosity drop-down list, keep the default selection of constant, and enter
a value of 0.001.
(e) Click Change/Create, and then click Yes in the dialog box prompting whether
you want to overwrite the definition of air.
(f) Click the Fluent Database... button in the Materials panel.
The Fluent Database Materials panel will open.

Modeling Cavitation
i. In the list of Fluent Fluid Materials, select water-vapor (h2o).

ii. Click Copy to copy the information for water vapor to your model.
iii. Close the Fluent Database Materials panel.
(g) Change the value of Density for water-vapor (h2o) to 0.02558.
(h) Change the value of Viscosity for water-vapor (h2o) to 1.26e-6.
(i) Click Change/Create button and close the Materials panel.
17-10
Modeling Cavitation
Step 4: Phases
1. Define the liquid water and water vapor phases that flow through the orifice.
(a) Specify liquid water as the primary phase.

ii. In the Primary Phase panel, enter liquid for the Name.
iii. Select water from the Phase Material drop-down list.
(b) Specify water vapor as the secondary phase.
ii. In the Secondary Phase panel, enter vapor for the Name.
iii. Select water-vapor from the Phase Material drop-down list.

Modeling Cavitation

1. Set the operating pressure to 0 pascal.
17-12
Modeling Cavitation

For this problem, you need to set the boundary conditions for two boundaries: the pressure
inlet (consisting of two boundary zones), and the pressure outlet. The pressure outlet is
the downstream boundary, opposite the pressure inlets.
1. Set the conditions for the pressure inlets (inlet-1, inlet-2).

For the multiphase mixture model, you will specify conditions for the mixture (i.e.,
conditions that apply to all phases) and also conditions that are specific to the
primary and secondary phases. In this tutorial, boundary conditions are needed for
the mixture and secondary phase only.
ii. Enter 500000 for the Gauge Total Pressure.
iii. Enter 449000 for the Supersonic/Initial Gauge Pressure.
If you choose to initialize the solution based on the pressure-inlet con-
ditions, the Supersonic/Initial Gauge Pressure will be used in conjunction
with the specified stagnation pressure (the Gauge Total Pressure) to com-
pute initial values according to the isentropic relations (for compressible
flow) or Bernoulli’s equation (for incompressible flow). Otherwise, in an
incompressible flow calculation, the Supersonic/Initial Gauge Pressure input
will be ignored by FLUENT. In this problem the velocity will be initialized
based on the difference between these two values.
iv. In the Direction Specification Method drop-down list, keep the default se-
lection of Normal to Boundary.
v. In the Turbulence Specification Method drop-down list, keep the default
selection of K and Epsilon.
vi. Under Turb. Kinetic Energy, enter 0.02.

Modeling Cavitation

i. In the Boundary Conditions panel, select vapor from the Phase drop-down
ii. Keep the default Volume Fraction of 0.

(c) Copy the boundary conditions defined for the first pressure inlet zone (inlet-1)
to the second one (inlet-2).
i. In the Boundary Conditions panel, select mixture from the Phase drop-down
list.
ii. In the Boundary Conditions panel, click Copy...
This will open the Copy BCs panel.
17-14
Modeling Cavitation
iii. Select inlet-1 in the From Zone list, and then select inlet-2 in the To Zones
list.
iv. Click Copy and close the panel.
2. Set the boundary conditions for the pressure outlet (outlet).

The turbulence conditions you input at the pressure outlet will be used only if flow
enters the domain through this boundary. You can set them equal to the inlet
values, as no flow reversal is expected at the pressure outlet. In general, however,
it is important to set reasonable values for these downstream scalar values, in case
flow reversal occurs at some point during the calculation.

Modeling Cavitation
ii. Under Gauge Pressure, enter 95000.

iii. Keep the default selection of K and Epsilon for the Turbulence Specification
Method.
iv. Set the Backflow Turb. Kinetic Energy to 0.02.
i. In the Boundary Conditions panel, select vapor from the Phase drop-down
ii. Retain the default Volume Fraction of 0.
17-16
Modeling Cavitation
Step 7: Solution
(a) Under Under-Relaxation Factors, set the under-relaxation factor for both Pres-
sure and Momentum to 0.4.
(b) Scroll down and set the under-relaxation factors for Turbulence Kinetic Energy,
Turbulence Dissipation Rate, and Turbulent Viscosity to 0.5.
(c) Keep the under-relaxation factor for Vaporization Mass to 1.
FLUENT’s new cavitation model follows a different numerical approach from
the previous one. In general it is more robust and gives more accurate results.
Typically, for more complex cases, with very high pressure drops or large liquid-
vapor density ratios, the under-relaxation factors may need to be reduced to
between 0.1 and 0.2. For the Vaporization Mass, it is advised to use a value of
0.1. But in this tutorial a value of 1 is used, since the solution diverges for a
value of 0.1.
(d) Under Discretization, select Linear in the Pressure drop-down list.
(e) Select SIMPLEC in the Pressure-Velocity Coupling drop-down list.

Modeling Cavitation

(a) Change all convergence criteria, except for vf-vapor, to 1e-5 for improved
accuracy.
(b) Select Plot under Options, and click OK.
17-18
Modeling Cavitation
3. Initialize the solution from either of the pressure inlet zones (inlet-1 or inlet-2).
(a) Select inlet-1 or inlet-2 in the Compute From drop-down list.

(b) Under Reference Frame, select Absolute.
(c) Click Init to initialize the solution.
4. Save the case file (cav.cas).


The solution will converge to within the specified criteria in approximately 1800
iterations.
6. Save the data file (cav.dat).


Modeling Cavitation
1. Plot the pressure in the orifice.
(a) Select Pressure... and Static Pressure in the drop-down lists under Contours of.
(c) Click Display.
Note the dramatic pressure drop at the flow restriction in Figure 17.3. Low
static pressure is the major factor to cause cavitation, though turbulence also
contributes to cavitation, due to the effect of pressure fluctuation and turbulent
diffusion, as will be shown in the following plots.
To make the view more realistic, you will need to mirror it across the center-
line.
17-20
Modeling Cavitation
4.99e+05
4.75e+05
4.50e+05
4.25e+05
4.00e+05
3.75e+05
3.51e+05
3.26e+05
3.01e+05
2.76e+05
2.51e+05
2.27e+05
2.02e+05
1.77e+05
1.52e+05
1.27e+05
1.03e+05
7.79e+04
5.31e+04
2.83e+04
3.54e+03
Contours of Static Pressure (mixture) (pascal)

FLUENT 6.2 (axi, segregated, mixture, ske)

Modeling Cavitation
2. Mirror the display across the centerline.

(a) Select symm-1 and symm-2 in the Mirror Planes list and click Apply.
4.99e+05
4.75e+05
4.50e+05
4.25e+05
4.00e+05
3.75e+05
3.51e+05
3.26e+05
3.01e+05
2.76e+05
2.51e+05
2.27e+05
2.02e+05
1.77e+05
1.52e+05
1.27e+05
1.03e+05
7.79e+04
5.31e+04
2.83e+04
3.54e+03

Figure 17.4: Mirrored View of Contours of Static Pressure
17-22
Modeling Cavitation
3. Plot the turbulent kinetic energy.

(a) Select Turbulence... and Turbulent Kinetic Energy in the drop-down lists under
Contours of.
(b) Click Display.
2.91e+01
2.77e+01
2.62e+01
2.48e+01
2.33e+01
2.18e+01
2.04e+01
1.89e+01
1.75e+01
1.60e+01
1.46e+01
1.31e+01
1.17e+01
1.02e+01
8.74e+00
7.29e+00
5.83e+00
4.37e+00
2.92e+00
1.46e+00
6.03e-03
Contours of Turbulence Kinetic Energy (k) (mixture) (m2/s2)

Figure 17.5: Contours of Turbulent Kinetic Energy
In this example, the grid used is fairly coarse. However, in cavitating flows the
pressure distribution is the dominant factor, and is not very sensitive to grid size.
4. Plot the volume fraction of water vapor.

(a) Select Phases... and Volume fraction in the drop-down lists under Contours of.
(b) Select vapor in the Phase drop-down list.
(c) Click Display.

Modeling Cavitation
1.00e-00
9.50e-01
9.00e-01
8.50e-01
8.00e-01
7.50e-01
7.00e-01
6.50e-01
6.00e-01
5.50e-01
5.00e-01
4.50e-01
4.00e-01
3.50e-01
3.00e-01
2.50e-01
2.00e-01
1.50e-01
1.00e-01
5.00e-02
0.00e+00
Contours of Volume fraction (vapor)

Figure 17.6: Contours of Vapor Volume Fraction
Note that the high turbulent kinetic energy region near the neck of the orifice (Fig-
ure 17.5) coincides with the highest volume fraction of vapor in Figure 17.6. This
indicates the correct prediction of a localized high phase change rate. The vapor
then gets convected downstream by the main flow.
Summary
This tutorial demonstrated how to set up and resolve a strongly cavitating pressure
driven flow through an orifice, using FLUENT’s multiphase mixture model with cavitation
effects. You learned how to set the boundary conditions for an internal flow. A steady-
state solution was calculated to simulate the formation of a vapor bubble in the neck
of the flow after the section restriction at the orifice. A more computationally-intensive
unsteady calculation is necessary to accurately simulate the irregular cyclic process of
bubble formation, growth, filling by water jet re-entry, and breakoff.
17-24
Tutorial 18. Using the Mixture and Eulerian Multiphase
Models
Introduction
This tutorial examines the flow of water and air in a tee junction. First you will solve
the problem using the less computationally-intensive mixture model, and then you will
turn to the more accurate Eulerian model. Finally, you will compare the results obtained
with the two approaches.
• Use the mixture model with slip velocities
• Use the Eulerian model
• Compare the results obtained with the two approaches
Prerequisites
Problem Description
This problem considers an air-water mixture flowing upwards in a duct and then splitting
in a tee-junction. The ducts are 25 mm in width, the inlet section of the duct is 125 mm
long, and the top and the side ducts are 250 mm long. The geometry and data for the
problem are shown in Figure 18.1.

Using the Mixture and Eulerian Multiphase Models
velocity inlet
water: v = - 0.31 m/s
air: v = - 0.45 m/s
pressure outlet
velocity inlet
water: air:
ρ=1000 kg/m3 ρ=1.2 kg/m3
µ=9e-4 kg/m-s µ=2e-5 kg/m-s
v=1.53 m/s v=1.6 m/s
vol frac=0.02
bubble diam=1
Setup and Solution

Preparation
1. Download mix_eulerian_multiphase.zip from the Fluent Inc. User Services Cen-
ter or copy it from the FLUENT documentation CD to your working directory (as
2. Unzip mix_eulerian_multiphase.zip.
tee.msh can be found in the /mix eulerian multiphase folder created after un-
zipping the file.
18-2
Step 1: Grid
1. Read the grid file (tee.msh).
2. Check the grid.

Grid −→Check

(a) Display the grid using the default settings (Figure 18.2).

Grid
Figure 18.2: The Grid in the Tee Junction
18-4
Step 2: Models
1. Keep the default settings for the 2D segregated steady-state solver.

2. Enable the multiphase mixture model with slip velocities.

(a) Select Mixture as the Model.
The panel will expand to show the inputs for the mixture model.
(b) Under Mixture Parameters, keep the Slip Velocity turned on.
Since there will be significant difference in velocities for the different phases,
you need to solve the slip velocity equation.
(c) Under Body Force Formulation, select Implicit Body Force.
This treatment improves solution convergence by accounting for the partial
equilibrium of the pressure gradient and body forces in the momentum equa-
tions. It is used when body forces are large in comparison to viscous and
convective forces, namely in VOF and mixture problems.
18-6

(b) Under k-epsilon Model, keep the default selection of Standard.
The standard k- model has been found to be quite effective in accurately re-
solving mixture problems when standard wall functions are used.
(c) Keep the default selection of Standard Wall Functions under Near-Wall Treat-
ment.
This problem does not require a particularly fine grid, and standard wall func-
tions will be used.


(b) Set the Gravitational Acceleration in the Y direction to -9.81 m/s2 .
18-8
Step 3: Materials
1. Copy liquid water from the materials database so that it can be used for the primary
phase.
(a) Click the Fluent Database... button in the Materials panel.
(b) In the list of Fluent Fluid Materials, select water-liquid (h2o<l>).

(c) Click Copy to copy the information for liquid water to your model.
(d) Close the Fluent Database Materials panel and the Materials panel.

Step 4: Phases
1. Define the liquid water and air phases that flow in the tee junction.
(a) Specify liquid water as the primary phase.

ii. In the Primary Phase panel, enter water under Name.

iii. Select water-liquid from the Phase Material drop-down list.
18-10
(b) Specify air as the secondary phase.

ii. In the Secondary Phase panel, enter air under Name.

iii. Retain air as the Phase Material.
iv. Set the Diameter to 0.001 m.
v. Click OK to close the panel.

2. Check the slip velocity formulation to be used.

(a) Click the Interaction... button in the Phases panel.
(b) In the Phase Interaction panel, click the Slip tab.

(c) Keep the default selection of manninen-et-al in the Slip Velocity drop-down list.
18-12

For this problem, you need to set the boundary conditions for three boundaries: the upper
and lower velocity inlets and the pressure outlet.
1. Set the conditions for the lower velocity inlet (velocity-inlet-4).

For the multiphase mixture model, you will specify conditions at a velocity inlet for
the mixture (i.e., conditions that apply to all phases) and also conditions that are
specific to the primary and secondary phases.
(a) Set the conditions at velocity-inlet-4 for the mixture.
ii. In the Turbulence Specification Method drop-down list, select Intensity and
Length Scale.
iii. Set Turbulence Intensity to 10% and Turbulence Length Scale to 0.025 m.

(b) Set the conditions for the primary phase.

ii. Keep the default Velocity Specification Method and Reference Frame.
iii. Set the Velocity Magnitude to 1.53.
(c) Set the conditions for the secondary phase.
i. In the Boundary Conditions panel, select air from the Phase drop-down list
and click Set....
iv. Set the Volume Fraction to 0.02.
18-14
2. Set the conditions for the upper velocity inlet (velocity-inlet-5).

(a) Set the conditions at velocity-inlet-5 for the mixture.
i. In the Boundary Conditions panel, select mixture in the Phase drop-down
Length Scale.
iii. Set Turbulence Intensity to 10% and Turbulence Length Scale to 0.025 m.
iii. Set the Velocity Magnitude to -0.31.
In this problem, outflow characteristics at the upper velocity inlet are as-
sumed to be known, and therefore imposed as a boundary condition.


and click Set....
iii. Set the Velocity Magnitude to -0.45.
iv. Set the Volume Fraction to 0.02.
18-16
3. Set the boundary conditions for the pressure outlet (pressure-outlet-3).

For the multiphase mixture model, you will specify conditions at a pressure outlet
for the mixture and for the secondary phase. There are no conditions to be set for
the primary phase.
The turbulence conditions you input at the pressure outlet will be used only if flow
enters the domain through this boundary. You can set them equal to the inlet
values, as no flow reversal is expected at the pressure outlet. In general, however,
it is important to set reasonable values for these downstream scalar values, in case
flow reversal occurs at some point during the calculation.
(a) Set the conditions at pressure-outlet-3 for the mixture.
Length Scale.
iii. Set the Backflow Turbulence Intensity to 10%.
iv. Set the Backflow Turbulence Length Scale to 0.025.
(b) Set the conditions for the secondary phase.
and click Set....
ii. Set the Backflow Volume Fraction to 0.02.

Step 6: Solution Using the Mixture Model

(a) Keep all default Under-Relaxation Factors.

(b) Under Discretization, select PRESTO! in the Pressure drop-down list.

(a) Turn on Plot.
18-18

(a) Click Init and close the panel.
4. Save the case file (tee.cas).


The solution will converge in approximately 660 iterations.
6. Save the case and data files (tee.cas and tee.dat).


Step 7: Postprocessing for the Mixture Solution

1. Display the pressure field in the tee (Figure 18.3).
(a) Select Pressure... and Static Pressure in the Contours of drop-down lists.
(c) Click Display.
18-20
2.35e+03
2.15e+03
1.96e+03
1.76e+03
1.57e+03
1.37e+03
1.18e+03
9.80e+02
7.85e+02
5.90e+02
3.96e+02
2.01e+02
5.63e+00
-1.89e+02
-3.84e+02
-5.79e+02
-7.74e+02
-9.69e+02
-1.16e+03
-1.36e+03
-1.55e+03

FLUENT 6.2 (2d, segregated, mixture, ske)

2. Display contours of velocity magnitude (Figure 18.4).

(a) Select Velocity... and Velocity Magnitude in the Contours of drop-down lists.
(b) Click Display.
2.24e+00
2.12e+00
2.01e+00
1.90e+00
1.79e+00
1.68e+00
1.57e+00
1.45e+00
1.34e+00
1.23e+00
1.12e+00
1.01e+00
8.95e-01
7.83e-01
6.71e-01
5.59e-01
4.47e-01
3.35e-01
2.24e-01
1.12e-01
0.00e+00
Contours of Velocity Magnitude (mixture) (m/s)

Figure 18.4: Contours of Velocity Magnitude
3. Display the volume fraction of air (Figure 18.5).

(a) Select Phases... and Volume fraction in the Contours of drop-down lists.
(b) Select air in the Phase drop-down list.
(c) Click Display.
In Figure 18.5, note the small bubble of air that separates at the sharp edge of the
horizontal arm of the tee junction, and the small layer of air that floats in the same
area above the water, marching towards the pressure outlet.
18-22
8.80e-01
8.36e-01
7.92e-01
7.48e-01
7.04e-01
6.60e-01
6.16e-01
5.72e-01
5.28e-01
4.84e-01
4.40e-01
3.96e-01
3.52e-01
3.08e-01
2.64e-01
2.20e-01
1.76e-01
1.32e-01
8.80e-02
4.40e-02
4.33e-13
Contours of Volume fraction (air)

Figure 18.5: Contours of Air Volume Fraction

Step 8: Setup and Solution for the Eulerian Model

You will use the solution obtained with the mixture model as an initial condition for the
calculation with the Eulerian model.
1. Turn on the Eulerian model.

(a) Under Model, select Eulerian.
18-24
2. Specify the drag law to be used for computing the interphase momentum transfer.
(b) In the Phase Interaction panel, keep the default selection of schiller-naumann in
the Drag Coefficient drop-down list.
Note: For this problem there are no parameters to be set for the individual phases,
other than those that you specified when you set up the phases for the mixture
model calculation. If you use the Eulerian model for a flow involving a granular
secondary phase, there are additional parameters that you need to set. There
are also other options in the Phase Interaction panel that may be relevant for
other applications.
See Section 24.8 of the User’s Guide for complete details on setting up an Eulerian
multiphase calculation.

3. Select the multiphase turbulence model.

(a) Under k-epsilon Multiphase Model, keep the default selection of Mixture.
The mixture turbulence model is applicable when phases separate, for stratified
(or nearly stratified) multiphase flows, and when the density ratio between
phases is close to 1. In these cases, using mixture properties and mixture
velocities is sufficient to capture important features of the turbulent flow.
See Chapter 24 of the User’s Guide for more information on turbulence models
for the Eulerian multiphase model.
18-26
4. Continue the solution by requesting 1200 additional iterations.

5. Save the case and data files (tee2.cas and tee2.dat).


Step 9: Postprocessing for the Eulerian Model

1. Display the pressure field in the tee (Figure 18.6).
2.40e+03
2.21e+03
2.02e+03
1.83e+03
1.64e+03
1.45e+03
1.27e+03
1.08e+03
8.87e+02
6.97e+02
5.08e+02
3.18e+02
1.29e+02
-6.04e+01
-2.50e+02
-4.39e+02
-6.29e+02
-8.18e+02
-1.01e+03
-1.20e+03
-1.39e+03

FLUENT 6.2 (2d, segregated, eulerian, ske)
2. Display contours of velocity magnitude for the water (Figure 18.7).

(a) In the Contours of drop-down lists, select Velocity... and water Velocity Magni-
tude.
Because the Eulerian model solves individual momentum equations for each
phase, you have the choice of which phase to plot solution data for.
(b) Click Display.
3. Display the volume fraction of air (Figure 18.8).

Note that the air bubble at the tee junction in Figure 18.8 is slightly different from
the one that you observed in the solution obtained with the mixture model (Fig-
ure 18.5). The Eulerian model generally offers better accuracy than the mixture
model, as it solves separate sets of equations for each individual phase, rather than
modeling slip velocity between phases.
See Chapter 24 of the User’s Guide for more information about the mixture and
Eulerian models.
18-28
2.37e+00
2.25e+00
2.13e+00
2.01e+00
1.90e+00
1.78e+00
1.66e+00
1.54e+00
1.42e+00
1.30e+00
1.19e+00
1.07e+00
9.48e-01
8.30e-01
7.11e-01
5.93e-01
4.74e-01
3.56e-01
2.37e-01
1.19e-01
0.00e+00
Contours of Velocity Magnitude (water) (m/s)

Figure 18.7: Contours of Water Velocity Magnitude
9.89e-01
9.40e-01
8.90e-01
8.41e-01
7.91e-01
7.42e-01
6.93e-01
6.43e-01
5.94e-01
5.44e-01
4.95e-01
4.45e-01
3.96e-01
3.46e-01
2.97e-01
2.47e-01
1.98e-01
1.48e-01
9.89e-02
4.95e-02
0.00e+00
Contours of Volume fraction (air)

Figure 18.8: Contours of Air Volume Fraction

Summary
This tutorial demonstrated how to set up and solve a multiphase problem using the
mixture model and the Eulerian model. You learned how to set boundary conditions for
the mixture and both phases. The solution obtained with the mixture model was used as a
starting point for the calculation with the Eulerian model. After completing calculations
with both models, you compared the results obtained with the two approaches.
18-30
Tutorial 19. Using the Eulerian Multiphase Model for
Granular Flow
Introduction
Mixing tanks are used to maintain solid particles or droplets of heavy fluids in suspension.
Mixing may be required to enhance reaction during chemical processing or to prevent
sedimentation. In this tutorial, you will use the Eulerian multiphase model to solve the
particle suspension problem. The Eulerian multiphase model solves momentum equations
for each of the phases, which are allowed to mix in any proportion.
• Use the granular Eulerian multiphase model
• Specify fixed velocities with a user-defined function (UDF) to simulate an impeller
• Solve a time-accurate transient problem
Prerequisites
Problem Description
The problem involves the transient startup of an impeller-driven mixing tank. The
primary phase is water, while the secondary phase consists of sand particles with a 111
micron diameter. The sand is initially settled at the bottom of the tank, to a level just
above the impeller. A schematic of the mixing tank and the initial sand position is shown
in Figure 19.1. The domain is modeled as 2D axisymmetric.

Using the Eulerian Multiphase Model for Granular Flow
.4446 m
.016 m
water
.4446 m
impeller
settled
sand .1728 m
.0864 m
bed .116 m
The fixed-values option will be used to simulate the impeller. Experimental data are used
to represent the time-averaged velocity and turbulence values at the impeller location.
This approach avoids the need to model the impeller itself. These experimental data are
provided in a user-defined function.
Setup and Solution

Preparation
1. Download eulerian_multiphase_granular.zip from the Fluent Inc. User Ser-
vices Center or copy it from the FLUENT documentation CD to your working
directory (as described in Tutorial 1).
2. Unzip eulerian_multiphase_granular.zip.
mixtank.msh and fix.c can be found in the /eulerian multiphase granular
folder created after unzipping the file.
19-2
Step 1: Grid
1. Read the grid file (mixtank.msh).
2. Check the grid.

Grid −→Check
3. Display the grid with default settings (Figure 19.2).

(a) Click Display in the Grid Display panel.
Grid
Extra: You can use the right mouse button to check which zone number corre-
sponds to each boundary. If you click the right mouse button on one of the
boundaries in the graphics window, its zone number, name, and type will be
them quickly.

4. Modify the grid colors.

(a) In the Grid Display panel, click the Colors... button.
This will open the Grid Colors panel. This panel allows you to control the colors
that are used to draw grids.
(b) In the Grid Colors panel, select Color by ID.

This will assign a different color to each zone in the domain, rather than to
each type of zone.
(c) In the Grid Display panel, click Display. (Figure 19.3).
The graphics display will be updated to show the grid.
Grid
FLUENT 6.2 (axi, segregated, lam)
Figure 19.3: Grid Display (Color by ID option)
19-4
5. Manipulate the grid display to show the full tank upright.

(a) Under Mirror Planes, select axis and click Apply.

The grid display will be updated to show both sides of the tank.
(b) Click Auto Scale.
This option is used to scale and center the current display without changing
its orientation (Figure 19.4).
Grid
Figure 19.4: Grid Display with Both Sides of the Tank

(c) Click on Camera... to rotate the view.

This will open the Camera Parameters panel.
i. With left mouse button drag the indicator of the dial in the counter-
clockwise direction till the upright view is displayed (Figure 19.5).
ii. Click Apply and close the Camera Parameters.
Grid
Figure 19.5: Grid Display of the Upright Tank
(d) Close the Views panel.

Note: When experimenting with different view manipulation techniques, you may
accidentally “lose” your geometry in the display. You can easily return to the
default (front) view by clicking on the Default button in the Views panel.
19-6
Step 2: Models
1. Specify a transient, axisymmetric model.

(b) Under Space, select Axisymmetric.
(c) Under Time, select Unsteady and click OK.
2. Enable the Eulerian multiphase model.

(a) Under Model, select Eulerian and click OK.

3. Turn on the k- turbulence model with standard wall functions.


(b) Retain Standard Wall Functions under Near-Wall Treatment.
This problem does not require a particularly fine grid, and standard wall func-
tions can be used.
19-8
(c) Under k-epsilon Multiphase Model, select the Dispersed model.

The dispersed turbulence model is applicable in this case because there is clearly
one primary continuous phase and the material density ratio of the phases is
about 2.5. Furthermore, the Stokes number is much less than 1. Therefore,
the particle’s kinetic energy will not depart significantly from that of the liquid.


(b) Set the Gravitational Acceleration in the X direction to -9.81 m/s2 .

Step 3: Materials
In this step, you will add liquid water to the list of fluid materials by copying it from the
Fluent Database for materials, and create a new material called sand.
1. Copy liquid water from the materials database so that it can be used for the primary
phase.
(a) Click the Fluent Database... button in the Materials panel.
(b) In the list of Fluent Fluid Materials, select water-liquid (h2o<l>).

(c) Click Copy to copy the information for liquid water to your model.
(d) Close the Fluent Database Materials panel.
19-10
2. Create a new material called sand.
(a) Enter sand in the Name text-entry box.

(b) Under Properties, enter 2500 kg/m3 as the Density.
(c) Remove the entry for Chemical Formula so the field is blank.
(d) Click Change/Create and close the Materials panel.
When you click Change/Create, a question dialog box will appear, asking you
if water-liquid should be overwritten. Click No to retain water-liquid and add
the new material, sand, to the list. The Materials panel will be updated to show
the new material name in the Fluid Materials list.

Step 4: Phases
1. Specify water as the primary phase.

(a) In the Phases panel, select phase-1 and click Set....
(b) Under Name, enter water.

(c) Select water-liquid from the Phase Material drop-down list and click OK.
2. Specify sand as the secondary phase.

19-12
(b) Under Name, enter sand.

(c) Select sand from the Phase Material drop-down list.
(d) Turn on the Granular option.
(e) Under Granular Temperature Model, retain Phase Property.
(f) Define the properties of the sand phase.
i. Enter 0.000111 as the Diameter.
ii. Select syamlal-obrien from the Granular Viscosity drop-down list.
iii. Select lun-et-al from the Granular Bulk Viscosity drop-down list.
iv. Enter 0.6 as the Packing Limit.

3. Specify the drag law to be used for computing the interphase momentum transfer.
(b) In the Phase Interaction panel, select gidaspow in the Drag Coefficient drop-
down list and click OK.
4. Close the Phases panel.
19-14

For this problem, there are no conditions to be specified on the outer boundaries. Within
the domain, there are three fluid zones, representing the impeller region, the region where
the sand is initially located, and the rest of the tank. There are no conditions to be
specified in the latter two zones, so you will need to set conditions only in the zone
representing the impeller.
As mentioned earlier, a UDF is used to specify the fixed velocities that simulate the
impeller. The values of the time-averaged impeller velocity components and turbulence
quantities are based on experimental measurement. The variation of these values may be
expressed as a function of radius, and imposed as polynomials according to:
variable = A1 + A2 r + A3 r2 + A4 r3 + ...
The order of polynomial to be used depends on the behavior of the function being fitted.
For this tutorial, the polynomial coefficients shown in Table 19.1 are provided in the UDF
fix.c.
Table 19.1: Impeller Profile Specifications
Variable A1 A2 A3 A4 A5 A6
u velocity -7.1357e-2 54.304 -3.1345e+3 4.5578e+4 -1.966e+5 –
v velocity 3.1131e-2 -10.313 9.5558e+2 -2.0051e+4 1.186e+5 –
kinetic energy 2.2723e-2 6.7989 -424.18 9.4615e+3 -7.725e+4 1.8410e+5
dissipation -6.5819e-2 88.845 -5.3731e+3 1.1643e+5 -9.120e+5 1.9567e+6
See the separate UDF Manual about setting up a UDF using the DEFINE PROFILE macro.
Note that, while this macro is usually used to specify a profile condition on a boundary
face zone, it is used in fix.c to specify the condition in a fluid cell zone. The arguments
of the macro have been changed accordingly.

1. Compile the UDF, fix.c, using the Interpreted UDFs panel.

Define −→ User-Defined −→ Functions −→Interpreted...
(a) Click Browse... and select the file, fix.c.

(b) Turn on the Display Assembly Listing option.
The Display Assembly Listing option displays the assembly language code in the
console window as the function compiles.
(c) Click Interpret to interpret the UDF.
Note: The name and contents of the UDF are stored in the case file when
you save the case file.
2. Set the conditions for the fluid zone representing the impeller (fix-zone).
Specify the conditions for water and sand separately using the UDF. For the mixture
(i.e., conditions that apply to all phases), the default conditions are acceptable.
(a) Set the water boundary conditions on fix-zone.
i. In the Boundary Conditions panel, select water under Phase and click Set....
19-16
ii. Turn on the Fixed Values option.

The panel will expand to show the related inputs.
iii. Set the following fixed values:
Parameter Value
Axial Velocity udf fixed u
Radial Velocity udf fixed v
Turbulence Kinetic Energy udf fixed ke
Turbulence Dissipation Rate udf fixed diss
(b) Set the conditions for sand on fix-zone.
i. In the Boundary Conditions panel, select sand under Phase and click Set....
ii. In the Fluid panel, turn on the Fixed Values option.
iii. Select udf fixed u for Axial Velocity and udf fixed v for Radial Velocity.

19-18
Step 6: Solution
(a) For the Under-Relaxation Factors, set Pressure to 0.5, Momentum to 0.2, and
Turbulent Viscosity to 0.8.
(b) Under Discretization, keep the default settings.

3. Initialize the solution using the default initial values.

4. Patch the initial sand bed configuration.


(a) Select sand in the Phase drop-down list.

(b) In the Variable list, select Volume Fraction.
(c) Select initial-sand in the Zones to Patch list and
(d) Set the Value to 0.56 and click Patch.
19-20
5. Set the time stepping parameters.

(a) Set the Time Step Size to 0.005.
(b) Under Iteration, set the Max Iterations per Time Step to 40 and click Apply.
6. Save the initial case and data files (mixtank.cas and mixtank.dat).
The problem statement is now complete. As a precaution, you should review the
impeller velocity fixes and sand bed patch after running the calculation for a single
time step. Since you are using a UDF for the velocity profiles, you need to perform
one time step in order for the profiles to be calculated and available for viewing.
7. Run the calculation for 0.005 seconds.
(a) Set the Number of Time Steps to 1 and click Iterate.
8. Check the initial velocities and sand volume fraction.
In order to display the initial fixed velocities in the fluid zone (fix-zone), you will
need to create a surface for this zone.

(a) Create a surface for fix-zone.

Surface −→Zone...
i. In the Zone list, select fix-zone and click Create.

The default name is the same as the zone name. FLUENT will automat-
ically assign the default name to the new surface when it is created. The
new surface will be added to the Surfaces list in the Zone Surface panel.
(b) Display the initial impeller velocities for water.
19-22
i. Select Velocity in the Vectors of drop-down list.

ii. Select water in the Phase drop-down list.
iii. Select Velocity... and Velocity Magnitude in the Color by drop-down lists.
iv. Select water in the Phase drop-down list below Color by.
v. In the Surfaces list, select fix-zone and click Display.
FLUENT will display the water velocity vector fixes at the impeller location,
as shown in Figure 19.6.
(c) Similarly, display the initial impeller velocities for sand (Figure 19.7).
i. Select sand in the Phase drop-down list below the Vectors of and Color by
drop-down list and click Display.

8.20e-01
7.79e-01
7.38e-01
6.97e-01
6.56e-01
6.15e-01
5.74e-01
5.33e-01
4.92e-01
4.51e-01
4.10e-01
3.69e-01
3.28e-01
2.87e-01
2.46e-01
2.05e-01
1.64e-01
1.23e-01
8.20e-02
4.10e-02
4.25e-05
water-velocity Colored By Velocity Magnitude (water) (m/s) (Time=5.0000e-03)

FLUENT 6.2 (axi, segregated, eulerian, ske, unsteady)
Figure 19.6: Initial Impeller Velocities for Water
8.25e-01
7.84e-01
7.42e-01
7.01e-01
6.60e-01
6.19e-01
5.77e-01
5.36e-01
4.95e-01
4.54e-01
4.12e-01
3.71e-01
3.30e-01
2.89e-01
2.47e-01
2.06e-01
1.65e-01
1.24e-01
8.25e-02
4.12e-02
0.00e+00
sand-velocity Colored By Velocity Magnitude (sand) (m/s) (Time=5.0000e-03)

Figure 19.7: Initial Impeller Velocities for Sand
19-24
(d) Display contours of sand volume fraction.

i. Select Phases... and Volume fraction in the Contours of drop-down lists.

ii. Select sand in the Phase drop-down list.
iii. Select Filled under Options.
iv. Click Display.
FLUENT will display the initial location of the settled sand bed, shown in Fig-
ure 19.8.

5.62e-01
5.34e-01
5.06e-01
4.78e-01
4.50e-01
4.22e-01
3.94e-01
3.65e-01
3.37e-01
3.09e-01
2.81e-01
2.53e-01
2.25e-01
1.97e-01
1.69e-01
1.41e-01
1.12e-01
8.43e-02
5.62e-02
2.81e-02
0.00e+00
Contours of Volume fraction (sand) (Time=5.0000e-03)

Figure 19.8: Initial Settled Sand Bed
9. Run the calculation for 1 second.

(a) Set the Number of Time Steps to 199.
(b) Click Iterate.
After 200 time steps have been computed (a total of 1 second of operation),
you will review the results before continuing.
10. Save the case and data files (mixtank1.cas and mixtank1.dat).
19-26
11. Examine the results of the calculation after 1 second.

(a) Display the velocity vectors in the whole tank for the water.
Remember to deselect fix-zone in the Surfaces list.

!
Figure 19.9 shows the water velocity vectors after 1 second of operation. The
circulation is confined to the region near the impeller, and has not yet had
time to develop in the upper portions of the tank.
8.20e-01
7.79e-01
7.38e-01
6.97e-01
6.56e-01
6.15e-01
5.74e-01
5.33e-01
4.92e-01
4.51e-01
4.10e-01
3.69e-01
3.28e-01
2.87e-01
2.46e-01
2.05e-01
1.64e-01
1.23e-01
8.20e-02
4.10e-02
4.25e-05
water-velocity Colored By Velocity Magnitude (water) (m/s) (Time=1.0000e+00)

Figure 19.9: Water Velocity Vectors after 1 Second
(b) Display the velocity vectors for the sand.

Figure 19.10 shows the sand velocity vectors after 1 second of operation. The
circulation of sand around the impeller is significant, but note that no sand
vectors are plotted in the upper part of the tank, where the sand is not yet
present.
(c) Display contours of sand volume fraction.
Notice that the action of the impeller draws clear fluid from above the originally
settled bed and mixes it into the sand. To compensate, the sand bed is lifted
up slightly. The maximum sand volume fraction has decreased as a result of
the mixing of water and sand.

8.25e-01
7.84e-01
7.42e-01
7.01e-01
6.60e-01
6.19e-01
5.77e-01
5.36e-01
4.95e-01
4.54e-01
4.12e-01
3.71e-01
3.30e-01
2.89e-01
2.47e-01
2.06e-01
1.65e-01
1.24e-01
8.25e-02
4.12e-02
0.00e+00
sand-velocity Colored By Velocity Magnitude (sand) (m/s) (Time=1.0000e+00)

Figure 19.10: Sand Velocity Vectors after 1 Second
5.48e-01
5.21e-01
4.94e-01
4.66e-01
4.39e-01
4.11e-01
3.84e-01
3.57e-01
3.29e-01
3.02e-01
2.74e-01
2.47e-01
2.19e-01
1.92e-01
1.65e-01
1.37e-01
1.10e-01
8.23e-02
5.48e-02
2.74e-02
0.00e+00
Contours of Volume fraction (sand) (Time=1.0000e+00)

Figure 19.11: Contours of Sand Volume Fraction after 1 Second
19-28
12. Continue the calculation for another 19 seconds.

(a) Set the Time Step Size to 0.01.
The initial calculation was performed with a very small time step size to sta-
bilize the solution. After the initial calculation, you can usually increase the
time step to speed up the calculation.
(c) Click Iterate.
The transient calculation will continue to 20 seconds.
13. Save the case and data files (mixtank20.cas and mixtank20.dat).

You will now examine the progress of the sand and water in the mixing tank after a total
of 20 seconds.The mixing tank has nearly, but not quite, reached a steady flow solution.
1. Display the velocity vectors for the water (Figure 19.12).

Figure 19.12 shows the water velocity vectors after 20 seconds of operation. The
circulation of water is now very strong in the lower portion of the tank, though
modest near the top.
8.91e-01
8.47e-01
8.02e-01
7.58e-01
7.13e-01
6.69e-01
6.24e-01
5.79e-01
5.35e-01
4.90e-01
4.46e-01
4.01e-01
3.57e-01
3.12e-01
2.68e-01
2.23e-01
1.79e-01
1.34e-01
8.96e-02
4.51e-02
5.63e-04
water-velocity Colored By Velocity Magnitude (water) (m/s) (Time=2.0000e+01)

Figure 19.12: Water Velocity Vectors after 20 Seconds
2. Display the velocity vectors for the sand (Figure 19.13).

Figure 19.13 shows the sand velocity vectors after 20 seconds of operation. The sand
has now been suspended much higher within the mixing tank, but does not reach the
upper region of the tank. The water velocity in that region is not sufficient to
overcome the gravity force on the sand particles.
19-30
8.86e-01
8.42e-01
7.98e-01
7.53e-01
7.09e-01
6.65e-01
6.20e-01
5.76e-01
5.32e-01
4.87e-01
4.43e-01
3.99e-01
3.55e-01
3.10e-01
2.66e-01
2.22e-01
1.77e-01
1.33e-01
8.86e-02
4.43e-02
0.00e+00
sand-velocity Colored By Velocity Magnitude (sand) (m/s) (Time=2.0000e+01)

Figure 19.13: Sand Velocity Vectors after 20 Seconds

3. Display contours of sand volume fraction (Figure 19.14).

2.47e-01
2.34e-01
2.22e-01
2.10e-01
1.97e-01
1.85e-01
1.73e-01
1.60e-01
1.48e-01
1.36e-01
1.23e-01
1.11e-01
9.87e-02
8.63e-02
7.40e-02
6.17e-02
4.93e-02
3.70e-02
2.47e-02
1.23e-02
0.00e+00
Contours of Volume fraction (sand) (Time=2.0000e+01)

Figure 19.14: Contours of Sand Volume Fraction after 20 Seconds
4. Display filled contours of static pressure for the mixture (Figure 19.15).
(a) Select mixture in the Phase drop-down list.
(b) Select Pressure... and Static Pressure in the Contours of drop-down lists.
(c) Click Display.
Figure 19.15 shows the pressure distribution after 20 seconds of operation.
The pressure field represents the hydrostatic pressure except for some slight
deviations due to the flow of the impeller near the bottom of the tank.
19-32
7.87e+01
4.14e+00
-7.04e+01
-1.45e+02
-2.19e+02
-2.94e+02
-3.69e+02
-4.43e+02
-5.18e+02
-5.92e+02
-6.67e+02
-7.41e+02
-8.16e+02
-8.90e+02
-9.65e+02
-1.04e+03
-1.11e+03
-1.19e+03
-1.26e+03
-1.34e+03
-1.41e+03
Contours of Static Pressure (mixture) (pascal) (Time=2.0000e+01)

Figure 19.15: Contours of Pressure after 20 Seconds
Summary
This tutorial demonstrated how to set up and solve a granular multiphase problem us-
ing the Eulerian multiphase model. The problem involved the 2D modeling of particle
suspension in a mixing tank, and postprocessing showed the near-steady-state behavior
of the sand in the mixing tank, under the assumptions made.

19-34
Tutorial 20. Modeling Solidification
Introduction
This tutorial illustrates how to set up and solve a problem involving solidification. In
this tutorial, you will learn how to:
• Define a solidification problem
• Define pull velocities for simulation of continuous casting
• Define a surface tension gradient for Marangoni convection
• Solve a solidification problem
Prerequisites
Problem Description
This tutorial demonstrates the setup and solution procedure for a fluid flow and heat
transfer problem involving solidification, namely the Czochralski growth process. The
geometry considered is a 2D axisymmetric bowl (shown in Figure 20.1), containing a liq-
uid metal. The bottom and sides of the bowl are heated above the liquidus temperature,
as is the free surface of the liquid. The liquid is solidified by heat loss from the crystal
and the solid is pulled out of the domain at a rate of 0.001 m/s and a temperature of
500 K. There is a steady injection of liquid at the bottom of the bowl with a velocity of
1.01 × 10−3 and a temperature of 1300 K. Material properties are listed in Figure 20.1.
Starting with an existing 2D mesh, the details regarding the setup and solution proce-
dure for the solidification problem are presented. The steady conduction solution for
this problem is computed as an initial condition. Then, the fluid flow is turned on to
investigate the effect of natural and Marangoni convection in an unsteady fashion.

Modeling Solidification
Figure 20.1: Solidification in Czochralski model
20-2
Setup and Solution

Preparation
1. Download solidification.zip from the Fluent Inc. User Services Center or copy
Tutorial 1).
2. Unzip solidification.zip.
solid.msh can be found in the /solidification folder created after unzipping the
file.
Step 1: Grid
1. Read the mesh file solid.msh.
As this mesh is read by FLUENT, messages appear in the console window reporting
the progress of the reading.
2. Check the grid.

Grid −→Check
FLUENT performs various checks on the mesh and reports the progress in the console
window. Pay particular attention to the minimum volume. Make sure this is a
positive number.


Grid
Figure 20.2: Graphics Display of Grid
20-4
Step 2: Models
1. Enable the modeling of axisymmetric swirl.
(a) Under Space, select Axisymmetric Swirl.

(b) Keep the default settings for everything else.

2. Define the solidification model.

Define −→ Models −→Solidification & Melting...
(a) Under Model, turn on Solidification/Melting.
The panel will expand to show the related inputs.
(b) Under Parameters, keep the default value for the Mushy Zone Constant.
The default value of 100000 is acceptable for most cases.
(c) Turn on Include Pull Velocities.
The panel will expand to show an additional input.
• Including the pull velocities accounts for the movement of the solidified
material as it is continuously withdrawn from the domain in the continu-
ous casting process.
• It is possible to have FLUENT compute the pull velocities during the cal-
culation, but this approach is computationally expensive, and is recom-
mended only if the pull velocities are strongly dependent on the location
of the liquid-solid interface. In this tutorial, you will patch values for the
pull velocities instead of having FLUENT compute them. See the User’s
Guide for more information.
• When you click OK in the Solidification and Melting panel, FLUENT will present
an Information dialog box telling you that available material properties have
changed for the solidification model. You will be setting properties later, so
you can simply click OK in the dialog box to acknowledge this information.
• FLUENT will automatically enable the energy calculation when you enable the
solidification model, so you need not visit the Energy panel.
20-6
3. Add the effect of gravity on the model.


(b) Set the Gravitational Acceleration in the X direction to -9.81 m/s2 .

Step 3: Materials
In this step, you will create a new material and specify its properties, including the melting
heat, solidus temperature, and liquidus temperature.
1. In the Name field, enter liquid-metal.
2. Specify the density as a function of temperature.

As shown in Figure 20.1, the density of the material is defined by a polynomial
function: ρ = 8000 − 0.1T .
(a) Select Polynomial in the Density drop-down list.
The Polynomial Profile panel will open.
20-8
(b) Increase the value of Coefficients to 2.

(c) Enter 8000 for coefficient 1 and -0.1 for coefficient 2.
When you click OK in the Polynomial Profile panel, a question dialog box will
appear, asking you if air should be overwritten. Click No to retain air and add
the new material, liquid-metal, to the list. The Materials panel will be updated
to show the new material name in the Fluid Materials list. You will need to
select liquid-metal in the Fluid Materials drop-down list to set the other material
properties.
3. Set the specific heat, Cp, to 680 J/kg-K.
4. Set the Thermal Conductivity to 30 W/m-K.
5. Set the Viscosity to 0.00553 kg/m-s.
6. Set the Melting Heat to 100000 J/kg.
7. Set the Solidus Temperature to 1100 K.
8. Set the Liquidus Temperature to 1200 K.
9. Click on Change/Create and close the Materials panel.


1. Set the boundary conditions for the fluid.
(a) Select liquid-metal in the Material Name drop-down list.
2. Set the boundary conditions for the velocity inlet.
(a) Set the Velocity Magnitude to 0.00101 m/s.

(b) Set the Temperature to 1300 K.
20-10
3. Set boundary conditions for the outlet.

Here, the solid is pulled out with a specified velocity, so a velocity inlet is used with
the velocities pointing outwards.
(a) In the Velocity Specification Method drop-down list, select Components.

The panel will change to show related inputs.
(b) Set the Axial-Velocity to 0.001 m/s.
(c) Set the Swirl Angular Velocity to 1 rad/s.
(d) Set the Temperature to 500 K.

4. Set the boundary conditions for the bottom wall.
(a) Select Temperature under Thermal Conditions.

20-12
5. Set the boundary conditions for the free surface.

The specified shear and Marangoni stress boundary conditions are useful in modeling
situations in which the shear stress (rather than the motion of the fluid) is known. A
free surface condition is an example of such a situation. In this case, the convection
is Marangoni stress driven and the shear stress is dependent on the surface tension,
which is a function of temperature.
(a) Specify the thermal conditions.

i. Select Convection under Thermal Conditions.
The panel will change to show related inputs.
ii. Set the Heat Transfer Coefficient to 100 W/m2 -K.
iii. Set the Free Stream Temperature to 1500 K.

(b) Specify the shear conditions.

i. Click the Momentum tab.
The wall motion and shear condition will be displayed.
ii. Under Shear Condition, select Marangoni Stress.

The Marangoni Stress condition allows you to specify the gradient of the
surface tension with respect to temperature at a wall boundary.
iii. Set the Surface Tension Gradient to -0.00036 N/m-K.
6. Set the boundary conditions for the side wall.

(a) Select Temperature under Thermal Conditions.
20-14
7. Set the boundary conditions for the solid wall.

(a) Specify the thermal conditions.
i. Select Temperature under Thermal Conditions.
ii. Set the Temperature to 500 K.
(b) Specify the wall motion.
i. Click the Momentum tab.
ii. Under Wall Motion, select Moving Wall.

The panel will expand to show additional parameters.
iii. Under Motion, select Rotational.
The panel changes to show the rotational speed.
iv. Under Speed, set the rotational velocity to 1.0 rad/s.

Step 5: Solution: Steady Conduction

In this step, you will disable the calculation of the flow and swirl velocity equations,
and calculate the conduction only. This steady-state solution will be used as the initial
condition for the time-dependent fluid flow and heat transfer calculation.

In this step, you will specify the discretization schemes to be used, and temporarily
turn off the calculation of the flow and swirl velocity equations.
(a) In the Equations list, deselect Flow and Swirl Velocity.

(b) Keep the default values for all Under-Relaxation Factors.
(c) Under Discretization, select PRESTO! for Pressure, First Order Upwind for Mo-
mentum, Swirl Velocity, and Energy.
(d) Select SIMPLE in the Pressure-Velocity Coupling drop-down list.
20-16

(a) Check that the value for Temperature is set to 300 K.

Since you are solving only the steady conduction problem, the initial values for
the pressure and velocities will not be used.
(b) Click on Init and Close the panel.

3. Define a custom field function for the swirl pull velocity.

You will use this field function to patch a variable value for the swirl pull velocity in
the next step. The swirl pull velocity is equal to Ωr, where Ω is the angular velocity
and r is the radial coordinate. Since Ω = 1 rad/s, you can simplify the equation to
simply r. In this example, the value of Ω is included for demonstration purposes.
(a) In the Field Functions drop-down lists, select Grid... and Radial Coordinate.
(b) Click the Select button.
radial-coordinate will appear in the Definition field. If you make a mistake,
click the DEL button on the calculator pad to delete the last item you added to
the function definition.
(c) Click the X button on the calculator pad.
(d) Click on 1.
(e) Enter omegar as the New Function Name.
(f) Click Define and close the panel.
If you wish to check the function definition, click on Manage... and select
omegar.
20-18
4. Patch the pull velocities.

As noted earlier, you will patch values for the pull velocities, rather than having
FLUENT compute them. Since the radial pull velocity is zero, you will patch just
the axial and swirl pull velocities.
(a) Specify the value of the axial pull velocity.

i. In the Variable list, select Axial Pull Velocity.
ii. Select fluid in the Zones to Patch list.
iii. Set the Value to 0.001 m/s.
iv. Click Patch.
(b) Specify the value of the swirl pull velocity.

i. In the Variable list, select Swirl Pull Velocity.

ii. Enable the Use Field Function option.
iii. Select omegar in the Field Function list.
iv. Click Patch.


(b) Click OK.
6. Save the initial case and data files (solid0.cas and solid0.dat).

20-20


(b) Select Temperature... and Static Temperature in the Contours of drop-down
lists.
(c) Click Display.

1.40e+03
1.36e+03
1.31e+03
1.27e+03
1.22e+03
1.18e+03
1.13e+03
1.09e+03
1.04e+03
9.95e+02
9.50e+02
9.05e+02
8.60e+02
8.15e+02
7.70e+02
7.25e+02
6.80e+02
6.35e+02
5.90e+02
5.45e+02
5.00e+02

FLUENT 6.2 (axi, swirl, segregated, lam)
Figure 20.3: Contours of Temperature for Steady Conduction Solution
The thickness of the mushy zone can be determined from the contours of tempera-
ture. The mushy zone is the region where the temperature is between the liquidus
temperature and solidus temperature.
9. Save the case and data files for the steady conduction solution (solid.cas and
solid.dat).
20-22
Step 6: Solution: Unsteady Flow and Heat Transfer

In this step, you will turn on time dependence and include the flow and swirl velocity
equations in the calculation. You will then solve the unsteady problem using the steady
conduction solution as the initial condition.
1. Enable a time-dependent solution.


(b) Under Unsteady Formulation, retain 1st-Order Implicit.

2. Enable the solution of the flow and swirl velocity equations.

(a) Select Flow and Swirl Velocity in the Equations list and keep the selection of
Energy.
Now all three items in the Equations list will be selected.
(b) Keep the default values for all Under-Relaxation Factors.
(c) Under Discretization, retain the settings for all parameters.
3. Save the initial case and data files (solid01.cas and solid01.dat).
4. Run the calculation for 2 time steps.

(a) Under Time, set the Time Step Size to 0.1 seconds.
(c) Under Iteration, retain the default value of 20 for Max Iterations per Time Step.
(d) Click Iterate.
20-24
5. Examine the results of the calculation after 0.2 seconds.

(a) Display filled contours of temperature (Figure 20.4).
i. Select Temperature... and Static Temperature in the Contours of drop-down
lists.
ii. Click Display.
1.40e+03
1.36e+03
1.31e+03
1.27e+03
1.22e+03
1.18e+03
1.13e+03
1.09e+03
1.04e+03
9.95e+02
9.50e+02
9.05e+02
8.60e+02
8.15e+02
7.70e+02
7.25e+02
6.80e+02
6.35e+02
5.90e+02
5.45e+02
5.00e+02
Contours of Static Temperature (k) (Time=2.0000e-01)

FLUENT 6.2 (axi, swirl, segregated, lam, unsteady)
Figure 20.4: Contours of Temperature at t = 0.2 s
The temperature contours show the gradient in temperature from the hot
walls on the left to the cooler zone on the right.
(b) Display contours of stream function (Figure 20.5).
i. Under Options, deselect Filled.
ii. Select Velocity... and Stream Function in the Contours of drop-down lists.
iii. Click Display.

2.18e-02
2.07e-02
1.97e-02
1.86e-02
1.75e-02
1.64e-02
1.53e-02
1.42e-02
1.31e-02
1.20e-02
1.09e-02
9.83e-03
8.74e-03
7.64e-03
6.55e-03
5.46e-03
4.37e-03
3.28e-03
2.18e-03
1.09e-03
0.00e+00
Contours of Stream Function (kg/s) (Time=2.0000e-01)

Figure 20.5: Contours of Stream Function at t = 0.2 s
As shown in Figure 20.5, the liquid is beginning to circulate in a large

eddy, driven by natural convection and Marangoni convection on the free
surface.
(c) Display contours of liquid fraction (Figure 20.6).
i. Under Options, select Filled.
ii. Select Solidification/Melting... and Liquid Fraction in the Contours of drop-
down lists.
iii. Click Display.
20-26
1.00e+00
9.50e-01
9.00e-01
8.50e-01
8.00e-01
7.50e-01
7.00e-01
6.50e-01
6.00e-01
5.50e-01
5.00e-01
4.50e-01
4.00e-01
3.50e-01
3.00e-01
2.50e-01
2.00e-01
1.50e-01
1.00e-01
5.00e-02
0.00e+00
Contours of Liquid Fraction (Time=2.0000e-01)

Figure 20.6: Contours of Liquid Fraction at t = 0.2 s
The liquid fraction contours show the current position of the melt front.
Note that in Figure 20.6, the mushy zone divides the liquid and solid
regions roughly in half.
6. Continue the calculation for 48 additional time steps.

After a total of 50 time steps have been completed, the elapsed time will be 5 seconds.

7. Examine the results of the calculation after 5 seconds.

(a) Display filled contours of temperature (Figure 20.7).
1.40e+03
1.36e+03
1.31e+03
1.26e+03
1.22e+03
1.18e+03
1.13e+03
1.08e+03
1.04e+03
9.95e+02
9.50e+02
9.05e+02
8.60e+02
8.15e+02
7.70e+02
7.25e+02
6.80e+02
6.35e+02
5.90e+02
5.45e+02
5.00e+02
Contours of Static Temperature (k) (Time=5.0000e+00)

Figure 20.7: Contours of Temperature at t = 5 s
As shown in Figure 20.7, the temperature contours are fairly uniform through
the melt front and solid material. The distortion of the temperature field due
to the recirculating liquid is also clearly evident.
In a continuous casting process, it is important to pull out the solidified ma-
terial at the proper time. If the material is pulled out too soon, it will not
have solidified; that is, it will still be in a mushy state. If it is pulled out too
late, it solidifies in the casting pool and cannot be pulled out in the required
shape. The optimal rate of pull can be determined from the contours of liquidus
temperature and solidus temperature.
20-28
(b) Display contours of stream function (Figure 20.8).

1.41e-01
1.34e-01
1.27e-01
1.20e-01
1.13e-01
1.05e-01
9.84e-02
9.14e-02
8.44e-02
7.73e-02
7.03e-02
6.33e-02
5.63e-02
4.92e-02
4.22e-02
3.52e-02
2.81e-02
2.11e-02
1.41e-02
7.03e-03
0.00e+00
Contours of Stream Function (kg/s) (Time=5.0000e+00)

Figure 20.8: Contours of Stream Function at t = 5 s
Note that the flow has developed more fully now, as compared with Figure 20.5
after 0.2 seconds. The main eddy, driven by natural convection and Marangoni
stress, dominates the flow.
To examine the position of the melt front and the extent of the mushy zone,
you will plot the contours of liquid fraction.

(c) Display filled contours of liquid fraction (Figure 20.9).
1.00e+00
9.50e-01
9.00e-01
8.50e-01
8.00e-01
7.50e-01
7.00e-01
6.50e-01
6.00e-01
5.50e-01
5.00e-01
4.50e-01
4.00e-01
3.50e-01
3.00e-01
2.50e-01
2.00e-01
1.50e-01
1.00e-01
5.00e-02
0.00e+00
Contours of Liquid Fraction (Time=5.0000e+00)

Figure 20.9: Contours of Liquid Fraction at t = 5 s
The introduction of liquid material at the left of the domain is balanced by the
pulling of the solidified material from the right. After 5 seconds, the equilib-
rium position of the melt front is beginning to be established.
8. Save the case and data files for the solution at 5 seconds (solid5.cas and solid5.dat).
20-30
Summary
In this tutorial, you studied the setup and solution for a fluid flow problem involving
solidification for the Czochralski growth process.
The solidification model in FLUENT can be used to model the continuous casting process
where a solid material is continuously pulled out from the casting domain. In this tutorial,
you patched a constant value and a custom field function for the pull velocities instead of
computing them. For cases where the pull velocity is not changing over the domain, this
approach is used as it is computationally less expensive than having FLUENT compute
the pull velocities during the calculation.
See Chapter 25 of the User’s Guide for more information about the solidification/melting
model.

20-32
Tutorial 21. Using the Eulerian Granular Multiphase Model
with Heat Transfer
Introduction
This tutorial examines the flow of air and a granular solid phase consisting of glass beads
in a hot gas fluidized bed, under uniform minimum fluidization conditions. The results
obtained for the local wall-to-bed heat transfer coefficient in FLUENT are compared with
the analytical results.
• Use the Eulerian granular model
• Use a user defined function (UDF) to specify a phase-specific velocity inlet profile
• Compare the results obtained with analytical results.
Prerequisites
Problem Description
This problem considers a hot gas fluidized bed in which air flows upwards through the
bottom of the domain and through an additional small orifice next to a heated wall. A
uniformly fluidized bed is examined, for the sake of comparison with analytical results
[1]. The geometry and data for the problem are shown in Figure 21.1.

Using the Eulerian Granular Multiphase Model with Heat Transfer
pressure outlet
101325 Pa
insulated wall heated wall

T = 373 K
0.598
volume
fraction
of solids
uniform velocity inlet orifice

u = 0.25 m/s u = 0.25 m/s
T = 293 K T = 293 K
Setup and Solution

Preparation
1. Download eulerian_granular_heat.zip from the Fluent Inc. User Services Cen-
ter or copy it from the FLUENT documentation CD to your working directory (as
2. Unzip eulerian_granular_heat.zip.
fluid-bed.msh and gasvel.c can be found in the /eulerian granular heat folder
created after unzipping the file.
3. Start the 2D double-precision version of FLUENT.
21-2
Step 1: Grid
1. Read the grid file (fluid-bed.msh).
2. Check the grid.

Grid −→Check
3. Display the grid with default settings (Figure 21.2).

(a) Click Display and close the panel.


Grid
FLUENT 6.2 (2d, dp, segregated, lam)
Figure 21.2: Grid Display of the Fluidized Bed
21-4
Step 2: Models
1. Enable the segregated unsteady solver.
(a) Under Time, select Unsteady and click OK.

2. Enable the Eulerian multiphase model for two phases.

(a) Select Eulerian as the Model and click OK.

4. Keep the default laminar model.

Experiments have shown negligible three-dimensional effects in the flow field for the
case modeled, suggesting very weak turbulent behavior.
21-6

(b) Set the Gravitational Acceleration in the Y direction to -9.81 m/s2 .

Step 3: Materials
1. Compile the user-defined function, gasvel.c, that will be used to define the thermal
conductivity for the gas and solid phase.
Define −→ User-Defined −→ Functions −→Compiled...
(a) Under Source Files, click Add...

(b) In the Select File panel, select the file gasvel.c and click OK.
(c) In the Compiled UDFs panel, click Build.
FLUENT will build a libudf directory and compile the UDF.
(d) Click OK in the dialog box that will appear.
(e) Click Load to open the directory and load the UDF.
21-8
2. Modify the properties for air, which will be used for the primary phase.
The properties used for air are modified to match data used by Kuipers et al. [1]
(a) Set the value of Density to 1.2.

(b) Set the value of Cp to 994.
(c) In the Thermal Conductivity drop-down list, select user-defined.
The User Defined Functions panel will automatically open.
(d) In the User Defined Functions panel, select conduct gas, and click OK.
(e) Click Change/Create.

3. Define a new fluid material for the granular phase (the glass beads).
(a) In the Name field, enter solids.

(b) Specify 2660 for the Density.
(c) Specify 737 for Cp.
(d) Retain user-defined in the Thermal Conductivity drop-down list, and click Edit...
(e) In the User Defined Functions panel, select conduct solid, and click OK.
(f) Click No in the dialog box asking if you want to overwrite air.
(g) Select solids under Fluent Fluid Materials.
(h) Click Change/Create.
21-10
Step 4: Phases
1. Define air as the primary phase.

(b) In the Primary Phase panel, enter air for the Name.
(c) Select air from the Phase Material drop-down list and click OK.
2. Define the solids (glass beads) as the secondary phase.


(b) In the Secondary Phase panel, enter solids for the Name.
(c) Select solids from the Phase Material drop-down list.
(d) Turn on Granular.
(e) Under Granular Temperature Model retain Phase Property option.
(f) Set the Diameter to 0.0005 m.
(g) In the Granular Viscosity drop-down list, select syamlal-obrien.
(h) In the Granular Bulk Viscosity drop-down list, select lun-et-al.
(i) Click OK.
21-12
3. Define the inter-phase interactions formulations to be used.

(b) In the Drag Coefficient drop-down list, select syamlal-obrien.

(c) Click the Heat tab, and select gunn in the Heat Transfer Coefficient drop-down
list.
The interphase heat exchange is simulated, using a drag coefficient, the default
restitution coefficient for granular collisions of 0.9, and a heat transfer coeffi-
cient. Granular phase lift is not very relevant in this problem, and in fact is
rarely used.
(d) Click OK.


For this problem, you need to set the boundary conditions for all boundaries.
1. Set the conditions for the lower velocity inlet (v uniform).

For the Eulerian multiphase model, you will specify conditions at a velocity inlet
that are specific to the primary and secondary phases.
(a) Set the conditions at v uniform for the primary phase.
i. In the Boundary Conditions panel, select air in the Phase drop-down list
and click Set....
iv. Set the Temperature to 293.
v. Click OK.
(b) Set the conditions at v uniform for the secondary phase.
i. In the Boundary Conditions panel, select solids from the Phase drop-down
iii. Keep the default value of 0 for the Velocity Magnitude.
v. Set the Granular Temperature to 0.0001.
vi. Keep the default value of 0 for the Volume Fraction.
vii. Click OK.
21-14
2. Set the conditions for the orifice velocity inlet (v jet).

(a) Set the conditions at v jet for the primary phase.
i. In the Boundary Conditions panel, select air in the Phase drop-down list
and click Set....
In order for a comparison with analytical results [1] to be meaningful, in
this simulation you will use a uniform value for the air velocity equal to
the minimum fluidization velocity at both inlets on the bottom of the bed.
v. Click OK.

(b) Set the conditions at v jet for the secondary phase.

iii. Keep the default value of 0 for the Velocity Magnitude.
v. Set the Granular Temperature to 0.0001.
vi. Keep the default value of 0 for the Volume Fraction.
vii. Click OK.
3. Set the boundary conditions for the pressure outlet (poutlet).

For the Eulerian granular model, you will specify conditions at a pressure outlet for
the mixture and for both phases.
The thermal conditions at the pressure outlet will be used only if flow enters the
domain through this boundary. You can set them equal to the inlet values, as no
flow reversal is expected at the pressure outlet. In general, however, it is important
to set reasonable values for these downstream scalar values, in case flow reversal
occurs at some point during the calculation.
(a) Set the conditions at poutlet for the mixture.
ii. Keep the default value of 0 for the Gauge Pressure.
iii. Click OK.
21-16

and click Set....
ii. Set the Backflow Total Temperature to 293.

iii. Click OK.
ii. Set the Backflow Total Temperature to 293.

iii. Set the Backflow Granular Temperature to 0.0001.
iv. Set the Backflow Volume Fraction to 0.
v. Click OK.

4. Set the boundary conditions for the heated wall (wall hot).
For the heated wall, you will set thermal conditions for the mixture, and momentum
conditions (zero shear) for both phases.
i. In the Boundary Conditions panel, select mixture from the Phase drop-down
ii. Select Temperature under Thermal Conditions, and input 373 for the Tem-
perature.
iii. Click OK.
21-18

and click Set....
ii. Click the Momentum tab.

iii. Select Specified Shear under Shear Condition (the panel will expand), and
keep the default values of 0 for the X-Component and Y-Component.
For the secondary phase, you will set the same conditions of zero shear as for
the primary phase.

5. Set the boundary conditions for the adiabatic wall (wall ins).
For the adiabatic wall, you will retain the default thermal conditions for the mixture
(zero heat flux), and set momentum conditions (zero shear) for both phases.
(a) Set the conditions for the primary phase.
and click Set....
ii. Select Specified Shear under Shear Condition (the panel will expand), and
keep the default values of 0 for the X-Component and Y-Component.
iii. Click OK.
(b) Set the conditions for the secondary phase.
For the secondary phase, you will set the same conditions of zero shear as for
the primary phase.
21-20
Step 6: Solution
(a) Set the under-relaxation factor for Pressure to 0.5.

(b) Set the under-relaxation factor for Momentum to 0.2.
(c) Set the under-relaxation factor for Volume Fraction to 0.5.
You will have to scroll the Under Relaxation Factors list to display the under-
relaxation factor for Volume Fraction.
(d) Keep all default Discretization schemes.
(e) Click OK.


3. Define a custom field function for the heat transfer coefficient.

First, define functions for the mixture temperature, and thermal conductivity, then
use these to define a function for the heat transfer coefficient.
(a) Define function t mix.
i. In the Field Functions drop-down list, select Temperature... and Static

Temperature.
ii. In the Phase drop-down list, select air, and click Select.
iii. Click the multiplication symbol in the calculator pad.
iv. In the Field Functions drop-down list, select Phases... and Volume Fraction.
v. In the Phase drop-down list, select air, and click Select.
vi. Click the addition symbol in the calculator pad.
vii. Repeat the steps above to add the term solids-temperature *
solids-vof.
viii. In the New Function Name field, enter t mix as the name of the defined
function.
ix. Click Define.
21-22
(b) Define function k mix.
i. In the Field Functions drop-down list, select Properties....

ii. In the Phase drop-down list, select air
iii. Set the property as Thermal Conductivity and click Select.
iv. Click the multiplication symbol in the calculator pad.
v. In the Field Functions drop-down list, select Phases... and Volume Fraction.
vi. In the Phase drop-down list, select air, and click Select.
vii. Click the addition symbol in the calculator pad.
viii. Repeat the steps above to add the term solids-thermal-
conductivity-lam * solids-vof.
ix. In the New Function Name field, enter k mix.
x. Click Define.
(c) Define function ave htc.
i. Click the subtraction symbol in the calculator pad.
ii. In the Field Functions drop-down list, select Custom Field Functions... and
k mix.
iii. Use the calculator pad and the Field Functions list as you did in the pre-
vious steps to complete the definition of the function as shown below.
−k mix × (t mix − 373)/(58.5 × 10(−6) )/80
iv. In the New Function Name field, enter ave htc.
v. Click Define.

4. Define the point surface in the cell next to the wall, on the plane y = 0.24.
(a) Under Coordinates, enter 0.28494 for x0, and 0.24 for y0.
(b) Enter y=0.24 for the New Surface Name.
(c) Click Create.
21-24
5. Define the surface monitor for the heat transfer coefficient.

(b) Enable Plot, Print, and Write for monitor-1.
(c) In the Every drop-down list, select Time Step, and click Define...
The Define Surface Monitor panel will open.
i. Enter htc in the Name field.

ii. In the Report Type drop-down list, select Facet Average.
iii. In the X Axis drop-down list, select Flow Time.
iv. Increase the Plot Window to 1.
v. In the Report of drop-down lists, select Custom Field Functions..., and
ave htc.
vi. In the Surfaces list, select y=0.24.
vii. Under File Name, enter htc-024 and click OK.
(d) Click OK in the Surface Monitors panel.


(a) Under Compute From, select all-zones.

(b) Retain default values and click Init.
7. Define an adaption register for the lower half of the fluidized bed.
This register is used to patch the initial volume fraction of solids in the next step.
Adapt −→Region...
21-26
(a) Under Input Coordinates, specify the Xmaximum value as 0.3 and the Ymaxi-
mum value as 0.5.
(b) Click Mark.
(c) Click Manage...
The Manage Adaption Registers panel will open.
(d) Under Registers, in the Manage Adaption Registers panel, select hexahedron-r0,
and click Display.
Adaption Markings (hexahedron-r0) (Time=0.0000e+00)

FLUENT 6.2 (2d, dp, segregated, eulerian, lam, unsteady)
Figure 21.3: Region Marked for Patching
After you define a region for adaption, it is good practice to display it to

visually verify that it encompasses the intended area.
8. Patch the initial volume fraction of solids in the lower half of the fluidized bed.
(a) In the Phase drop-down list, select solids.
(b) In the Variable drop-down list, select Volume Fraction.
(c) In the Value field, enter 0.598.
(d) In the Registers to Patch list, select hexahedron-r0.
(e) Click Patch.

At this point, it is good practice to display contours of the variable you just patched,
to ensure that the desired field was obtained.
9. Display contours of Volume Fraction of solids (Figure 21.4).

5.98e-01
5.68e-01
5.38e-01
5.08e-01
4.78e-01
4.49e-01
4.19e-01
3.89e-01
3.59e-01
3.29e-01
2.99e-01
2.69e-01
2.39e-01
2.09e-01
1.79e-01
1.50e-01
1.20e-01
8.97e-02
5.98e-02
2.99e-02
0.00e+00
Contours of Volume fraction (solids) (Time=0.0000e+00)

Figure 21.4: Initial Volume Fraction of Granular Phase (solids).
10. Save the case file (fluid-bed.cas).

21-28
11. Set a time step size of 0.00025 s and run the calculation for 7000 time steps.
The plot of the value of the mixture-averaged heat transfer coefficient in the cell
next to the heated wall versus time is in excellent agreement with results published
for the same case [1].
3000.0000
2500.0000
2000.0000
Vertex 1500.0000
Average
ave_htc
1000.0000
500.0000
0.0000
0.0000 0.2000 0.4000 0.6000 0.8000 1.0000 1.2000 1.4000 1.6000 1.8000
Flow Time
Convergence history of ave_htc on y=0.24 (in SI units) (Time=1.7500e+00)

Figure 21.5: Plot of Mixture-Averaged Heat Transfer Coefficient in the Cell Next to the
Heated Wall Versus Time
12. Save the case and data files (fluid-bed.cas and fluid-bed.dat).
Extra: If you decide to read in the case file that is provided for this tutorial on the
documentation CD, you will need to compile the UDF associated with this tutorial
in your working directory. This is necessary because FLUENT will expect to find
the correct UDF libraries in your working directory when reading the case file.

1. Display the pressure field in the fluidized bed (Figure 21.6).
(a) Select Pressure... and Static Pressure in the Contours of drop-down lists.
(c) Click Display.
Note the build-up of static pressure in the granular phase.
2. Display the volume fraction of solids (Figure 21.7).
(a) Select Phases... and Volume fraction in the Contours of drop-down lists.
(b) Select solids in the Phase drop-down list.
(c) Click Display.
(d) Zoom in to show the contours close to the region where the change in volume
fraction is the greatest.
Note that the region occupied by the granular phase has expanded slightly, as a result
of fluidization.
21-30
7.79e+03
7.40e+03
7.01e+03
6.62e+03
6.23e+03
5.84e+03
5.45e+03
5.06e+03
4.67e+03
4.29e+03
3.90e+03
3.51e+03
3.12e+03
2.73e+03
2.34e+03
1.95e+03
1.56e+03
1.17e+03
7.79e+02
3.89e+02
-8.77e-02
Contours of Static Pressure (mixture) (pascal) (Time=1.7500e+00)

6.29e-01
5.98e-01
5.67e-01
5.35e-01
5.04e-01
4.72e-01
4.41e-01
4.09e-01
3.78e-01
3.46e-01
3.15e-01
2.83e-01
2.52e-01
2.20e-01
1.89e-01
1.57e-01
1.26e-01
9.44e-02
6.29e-02
3.15e-02
0.00e+00
Contours of Volume fraction (solids) (Time=1.7500e+00)

Figure 21.7: Contours of Volume Fraction of Solids

Summary
This tutorial demonstrated how to set up and solve a granular multiphase problem with
heat transfer, using the Eulerian model. You learned how to set boundary conditions
for the mixture and both phases. The solution obtained is in excellent agreement with
analytical results from Kuipers et al. [1].
References
1. J. A. M. Kuipers, W. Prins, and W. P. M. Van Swaaij “Numerical Calculation
of Wall-to-Bed Heat Transfer Coefficients in Gas-Fluidized Beds”, Department of
Chemical Engineering, Twente University of Technology, in AIChE Journal, July
1992, Vol. 38, No. 7.
21-32
Tutorial 22. Postprocessing
Introduction
This tutorial demonstrates the postprocessing capabilities of FLUENT using a 3D model
of a flat circuit board with a heat generating electronic chip mounted on it. The flow
over the chip is laminar and involves conjugate heat transfer.
The heat transfer involves conduction in the chip and conduction and convection in the
surrounding fluid. The physics of conjugate heat transfer such as this, is common in many
engineering applications, including the design and cooling of electronic components.
In this tutorial, you will read the case and data files (without doing the calculation) and
perform a number of postprocessing exercises. In the process, you will learn how to:
• Add lights to the display at multiple locations
• Create surfaces for the display of 3D data
• Display filled contours of temperature on several surfaces
• Display velocity vectors
• Mirror a display about a symmetry plane
• Create animations
• Display results on successive slices of the domain
• Display pathlines
• Plot quantitative results
• Overlay and explode a display
• Annotate the display
Prerequisites

Postprocessing
Problem Description
The problem considered is shown schematically in Figure 22.1. The configuration consists
of a series of side-by-side electronic chips, or modules, mounted on a circuit board. Air
flow, confined between the circuit board and an upper wall, cools the modules. To take
advantage of the symmetry present in the problem, the model will extend from the middle
of one module to the plane of symmetry between it and the next module.
As shown in the figure, each half-module is assumed to generate 2.0 Watts and to have a
bulk conductivity of 1.0 W/m2 -K. The circuit board conductivity is assumed to be one
order of magnitude lower: 0.1 W/m2 -K. The air flow enters the system at 298 K with
a velocity of 1 m/s. The Reynolds number of the flow, based on the module height, is
about 600. The flow is therefore treated as laminar.
Symmetry
Planes
Top Wall
Externally Cooled
Bottom Wall
Externally Cooled
Electronic Module (one half)
Air Flow k = 1.0 W/m2-K
1.0 m/s Q = 2.0 Watts
298 K
Circuit Board
k = 0.1 W/m2-K
Setup and Solution

Preparation
1. Download postprocess.zip from the Fluent Inc. User Services Center or copy it
from the FLUENT documentation CD to your working directory (as described in
Tutorial 1).
2. Unzip postprocess.zip.
chip.cas and chip.dat can be found in the /postprocess folder created after
unzipping the file.
22-2
Postprocessing
Step 1: Grid Display

1. Read in the case and data files, chip.cas and chip.dat.
When you select the case file, the data file is read automatically.

(a) Under Options, retain Edges.

(b) In the Surfaces list, deselect all the surfaces and select board-top and chip.
To deselect all surfaces click on the far-right unshaded button at the top of the
Surfaces list, and then select the desired surfaces in the Surfaces list.
(c) Click Colors... and select Color by ID under options.
(d) Click Reset Colors and close the panel.
(e) Click Display.
3. Rotate and zoom the view.

(a) Use the left mouse button to rotate the view.
(b) Use the middle mouse button to zoom the view until you obtain an enlarged
isometric display of the circuit board in the region of the chip, as shown in
Figure 22.2.

Postprocessing
Y X
Z
Grid
Figure 22.2: Grid Display of the Chip and Board Surfaces
Extra: You can click the right mouse button on one of the boundaries displayed
in the graphics window and its zone number, name, and type will be printed
in the console window. This feature is especially useful when you have several
zones of the same type and you want to distinguish between them.
4. Display the grid faces (filled-grid).

(a) Under Options, deselect Edges, select Faces, and click Display.
The surfaces run together with no shading to separate the chip from the board.
22-4
Postprocessing
Step 2: Adding Lights

1. Add default lighting effects.
The default light settings add a white light at the position (1, 1, 1). The default
light is defined in the Lights panel by the Light ID 0 with Direction vectors (X, Y, Z)
as (1, 1,1).
(a) Under Lighting Attributes, turn on the Lights On option.

(b) Click Apply and close the panel.
Shading is added to the surface grid display (Figure 22.3).

Postprocessing
Y
X
Z
Grid
Figure 22.3: Graphics Display with Default Lighting
22-6
Postprocessing
2. Adding lights in specific directions.

Display −→Lights...
You can also open the Lights panel by clicking on the Lights... button in the Display
Options panel.
(a) Add a light at (-1,1,1).

i. Increase the Light ID to 1 and turn on the Light On option.
ii. Set X, Y, and Z to -1, 1, and 1.
iii. Turn on the Headlight On option.
iv. Click Apply.
(b) Repeat this procedure to add a second light (Light ID=2) at (-1,1,-1).
The result is a more softly shaded display (Figure 22.4).
The Headlight On option provides constant lighting effect from a light source directly
in front of the model, in the direction of the view. You can turn off the headlight
by deselecting the Headlight On option (Figure 22.5).
Extra: You can use the left mouse button to rotate the ball in the Active Lights window
to gain a perspective view on the relative locations of the lights that are currently
active, and see the shading effect on the ball at the center.
You can also change the color of one or more of the lights by typing the name of a
color in the Color field or moving the Red, Green, and Blue sliders.

Postprocessing
Y
X
Z
Grid
Figure 22.4: Display with Additional Lighting
Y
X
Z
Grid
Figure 22.5: Display with Additional Lighting: Headlight Off
22-8
Postprocessing
Step 3: Creating Isosurfaces

To display results in a 3D model, you will need surfaces on which the data can be displayed.
FLUENT creates surfaces for all boundary zones automatically. In the case file that you
have read, several of these surfaces have been renamed. Examples are board-sym and
board-ends, which correspond to the side and end faces of the circuit board.
You can define additional surfaces for viewing the results, for example, a plane in Carte-
sian space. In this exercise, you will create a horizontal plane cutting through the middle
of the module with a y value of 0.25 inches. You can use this surface to display the
temperature and velocity fields.
1. Create a surface of constant y coordinate.

(a) In the Surface of Constant drop-down lists, select Grid... and Y-Coordinate.
(b) Click Compute.
The Min and Max fields will display the y extents of the domain.
(c) Enter 0.25 under Iso-Values.
(d) Enter y=0.25in under New Surface Name.
(e) Click Create, and Close the panel.

Postprocessing
Step 4: Contours
1. Display filled contours of temperature on the symmetry plane (Figure 22.6).

(b) Select Temperature... and Static Temperature in the Contours Of drop-down
lists.
(c) In the Surfaces list, select board-sym, chip-sym, and fluid-sym.
(d) Click Display.
The temperature contour will be displayed.
(e) Rotate and zoom the display using the left and middle mouse buttons, respec-
tively, to obtain the view shown in Figure 22.6.
Hint: If the display disappears from the screen at any time, or if you are having
difficulty manipulating it with the mouse, you can open the Views panel from
the Display pull-down menu and use the Default button to reset the view. Al-
ternatively, you can revert to a previous graphics display using the keyboard
shortcut <Ctrl>-L.
The peak temperatures in the chip appear where the heat is generated, along with
the higher temperatures in the wake where the flow is recirculating.
22-10
Postprocessing
4.09e+02
4.03e+02
3.98e+02
3.92e+02
3.87e+02
3.81e+02
3.76e+02
3.70e+02
3.64e+02
3.59e+02
3.53e+02
3.48e+02
3.42e+02
3.37e+02
3.31e+02
3.26e+02
3.20e+02
3.15e+02
3.09e+02 Y
3.04e+02 X
Z
2.98e+02

Figure 22.6: Filled Contours of Temperature on the Symmetry Surfaces
2. Display filled contours of temperature on the plane, y=0.25in (Figure 22.7).

(a) Under Surfaces, deselect all surfaces, select y=0.25in, and click Display.
(b) Change the location of the colormap in the graphics display window.

Postprocessing
i. Under Captions, select Bottom in the Colormap Alignment drop-down list.

ii. Deselect Axes Visible and click Apply.
(c) Zoom the display using the middle mouse button to obtain the view shown in
Figure 22.7.
2.98e+02 3.15e+02 3.31e+02 3.48e+02 3.64e+02 3.81e+02 3.98e+02 4.09e+02

Figure 22.7: Temperature Contours on the Surface, y = 0.25 in.
In the contour display (Figure 22.7), the high temperatures in the wake of
the module are clearly visible. You can also display other quantities using the
Contours panel, such as velocity magnitude or pressure.
22-12
Postprocessing
Step 5: Velocity Vectors

Velocity vectors provide an excellent visualization of the flow around the module, depicting
details of the wake structure.
1. Display velocity vectors on the symmetry plane through the module centerline
(Figure 22.8).
(a) In the Surfaces list, select fluid-sym.

(b) Set the Scale Factor to 1.9.
(c) Click Display.
(d) In the Display Options panel, turn on Axes Visible and set the Color Map Align-
ment to Left and click Apply.
(e) Rotate and zoom the display to observe the vortex near the stagnation point
and in the wake of the module (Figure 22.8).

Postprocessing
1.41e+00
1.34e+00
1.27e+00
1.20e+00
1.13e+00
1.06e+00
9.89e-01
9.20e-01
8.50e-01
7.81e-01
7.11e-01
6.41e-01
5.72e-01
5.02e-01
4.33e-01
3.63e-01
2.94e-01
2.24e-01
1.54e-01 Y
8.49e-02
1.53e-02 Z X

Figure 22.8: Velocity Vectors in the Module Symmetry Plane
Note: The vectors in Figure 22.8 are shown without arrowheads. You can modify
the arrow style in the Vectors panel by selecting a different option from the
Style drop-down list.
Extra: If you want to decrease the number of vectors displayed, you can increase
the Skip factor to a non-zero value.
2. Plot velocity vectors in the horizontal plane intersecting the module (Figure 22.9).
After plotting the vectors, you will enhance the view by mirroring the display about
the module centerline and displaying the module surfaces.
(a) Deselect all surfaces by clicking the unshaded icon to the right of Surfaces.
(b) In the Surfaces list, select y=0.25in.
(c) Set the Scale to 3.8.
(d) Under Options, select Draw Grid to display the module surfaces.
22-14
Postprocessing
i. Under Options, select Faces.

ii. In the Surfaces list, select board-top and chip.
iii. Click Colors... to set the color of the surfaces.
This will open the Grid Colors panel.
A. Under Options select Color by Type.

B. In the Types list, select wall.
C. In the Colors list, select light blue, and close the panel.
iv. In the Grid Display panel, click Display and close the panel.
(e) In the Vectors panel, click Display.
(f) Rotate the display with the mouse to obtain the view shown in Figure 22.9.

Postprocessing
1.41e+00
1.34e+00
1.27e+00
1.20e+00
1.13e+00
1.06e+00
9.89e-01
9.20e-01
8.50e-01
7.81e-01
7.11e-01
6.41e-01
5.72e-01
5.02e-01
4.33e-01
3.63e-01
2.94e-01
2.24e-01
1.54e-01 Y
8.49e-02 X
1.53e-02
Z

Figure 22.9: Velocity Vectors Intersecting the Surface
3. Mirror the view about the chip symmetry plane (Figure 22.10).
(a) In the Mirror Planes list, select symmetry-18.

Note: This zone is the centerline plane of the module, and its selection will
create a mirror of the entire display about the centerline plane.
(b) Click Apply.
The display will be updated in the graphics window (Figure 22.10).
22-16
Postprocessing
1.41e+00
1.34e+00
1.27e+00
1.20e+00
1.13e+00
1.06e+00
9.89e-01
9.20e-01
8.50e-01
7.81e-01
7.11e-01
6.41e-01
5.72e-01
5.02e-01
4.33e-01
3.63e-01
2.94e-01
2.24e-01
1.54e-01 Y
8.49e-02 X
1.53e-02 Z

Figure 22.10: Velocity Vectors after Mirroring

Postprocessing
Step 6: Animation
Using FLUENT, you can animate the solution and also a scene. For information on
animating the solution, see Tutorial 11, Steps 9 and 10. In this tutorial you will animate
a scene between two static views of the graphics display.
Display the surface temperature distribution on the module and the circuit board by select-
ing the corresponding boundaries. Create the key frames and view the transition between
the key frames, dynamically, using the animation feature.
1. Display filled contours of surface temperature on the board-top and chip surfaces.
(Figure 22.11).

(b) Select Temperature... and Static Temperature in the Contours Of drop-down
lists.
(c) Deselect all surfaces by clicking the unshaded icon to the right of Surfaces.
(d) In the Surfaces list, select board-top and chip.
(e) Click Display.
(f) Zoom the display as needed to obtain the view shown in Figure 22.11.
22-18
Postprocessing
4.09e+02
4.03e+02
3.98e+02
3.92e+02
3.87e+02
3.81e+02
3.76e+02
3.70e+02
3.64e+02
3.59e+02
3.53e+02
3.48e+02
3.42e+02
3.37e+02
3.31e+02
3.26e+02
3.20e+02
3.15e+02
3.09e+02 Y
3.04e+02 X
2.98e+02 Z

Figure 22.11: Filled Temperature Contours on the Chip and Board Top Surfaces
Figure 22.11 shows the high temperatures on the downstream portions of the module
and the relatively localized heating of the circuit board around the module.
2. Create the key frames by changing the point of view.

Display −→Scene Animation...
You will use the current display (Figure 22.11) as the starting view for the anima-
tion (Frame = 1).

Postprocessing
(a) Under Key Frames, click Add.

This will store the current display as Key-1.
(b) Zoom the view to focus on the module region.
(c) Under Key Frames, change the Frame number to 10.
(d) Click Add.
This will store the new display as Key-10.
The zoomed view will be the tenth keyframe of the animation, with intermediate
displays (2 through 9) to be filled in during the animation.
(e) Rotate the view and un-zoom the display so that the downstream side of the
module is in the foreground, as shown in Figure 22.12.
(f) Change the Frame number to 20.
(g) Click Add.
This will store the new display as Key-20.
3. View the scene animation.
(a) Click on the “play” arrow button ( ) (second from the right in the row of
playback buttons) in the Playback section of the Animate panel.
While effective animation is best conducted on “high-end” graphics workstations,
you can view scene animations on any workstation. If the graphics display speed is
slow, the animation playback will take some time and will appear choppy, with the
redrawing very obvious. On fast graphics workstations, the animation will appear
smooth and continuous and will provide an excellent visualization of the display
from a variety of spatial orientations. On many machines, you can improve the
22-20
Postprocessing
4.09e+02
4.03e+02
3.98e+02
3.92e+02
3.87e+02
3.81e+02
3.76e+02
3.70e+02
3.64e+02
3.59e+02
3.53e+02
3.48e+02
3.42e+02
3.37e+02
3.31e+02
3.26e+02
3.20e+02
3.15e+02
3.09e+02 Y
3.04e+02
X
2.98e+02Z

Figure 22.12: Filled Temperature Contours on the Chip and Board Top Surfaces
smoothness of the animation by turning on the Double Buffering option in the Display
Options panel.
Note: You can also make use of FLUENT’s animation tools for transient cases as demon-
strated in Tutorial 4.
Extra: You can change the Playback mode if you want to “auto repeat” or “auto reverse”
the animation. When you are in either of these Playback modes, you can click on
the “stop” button (square) to stop the continuous animation.

Postprocessing
Step 7: Pathlines
Pathlines are the lines traveled by neutrally buoyant particles in equilibrium with the fluid
motion. Pathlines are an excellent tool for visualization of complex three-dimensional
flows. In this example, you will use pathlines to examine the flow around and in the wake
of the module.
1. Create a rake from which the pathlines will emanate.

(a) In the Type drop-down list, select Rake.

A rake surface consists of a specified number of points equally spaced between
two specified endpoints. A line surface (the other option in the Type list) is
a line that includes the specified endpoints and extends through the domain;
data points on a line surface will not be equally spaced.
(b) Keep the default of 10 for the Number of Points along the rake.
This will generate 10 pathlines.
(c) Under End Points, enter the coordinates of the line, using a starting coordinate
of (1.0, 0.105, 0.07) and an ending coordinate of (1.0, 0.25, 0.07), as shown in
the panel above.
This will define a vertical line in front of the module, about halfway between
the centerline and edge.
22-22
Postprocessing
(d) Enter pathline-rake for the New Surface Name.

You will refer to the rake by this name when you plot the pathlines.
(e) Click Create, and Close the panel.
2. Draw the pathlines (Figure 22.13).

(a) In the Release From Surfaces list, select pathline-rake.

(b) Set the Step Size to 0.001 inch and the number of Steps to 6000.
Note: A simple rule of thumb to follow when you are setting these two pa-
rameters is that if you want the particles to advance through a domain of
length L, the Step Size times the number of Steps should be approximately
equal to L.
(c) Set Path Coarsen to 5.
Coarsening the path line simplifies the plot and reduces the plotting time. The
coarsening factor specified under Path Coarsen indicates the interval at which
the points are plotted for a given path line in any cell.
(d) Under Options, select Draw Grid.

Postprocessing
(e) In the Surfaces list, select board-top and chip.

These surfaces should already be selected from the earlier exercise where the
grid was displayed with velocity vectors, Step 5: Velocity Vectors.
(f) Under Options, check that Faces is selected, and then Close the panel.
(g) In the Path Lines panel, click Display.
The pathlines will be drawn on the surface.
(h) Rotate the display so that the flow field in front and in the wake of the chip
is visible, as shown in Figure 22.13.
9.00e+00
8.55e+00
8.10e+00
7.65e+00
7.20e+00
6.75e+00
6.30e+00
5.85e+00
5.40e+00
4.95e+00
4.50e+00
4.05e+00
3.60e+00
3.15e+00
2.70e+00
2.25e+00
1.80e+00
1.35e+00
9.00e-01 Y
4.50e-01 X
0.00e+00 Z
Path Lines Colored by Particle ID Dec 16, 2002

Figure 22.13: Pathlines Display
3. Write the path lines to a file.

(a) In the Path Lines panel, under Options, select Write to File.
(b) Click Write....
The Display button changes toWrite button when you turn on the Write to File
option.
(c) In the Select File panel, enter chip-pathline as the Fieldview File name and
click OK.
FLUENT saves the file in Fieldview format (.fvp extension).
22-24
Postprocessing
4. Display pathlines as spheres.

(a) In the Release From Surfaces list, keep the selection of pathline-rake.
(b) Select sphere from the Style drop-down list.
(c) Click the Style Attributes button.
This displays the Path Style Attributes panel.
i. Set the Diameter to 0.0005.

Postprocessing
ii. Click OK to close the panel.

(d) Set the Step Size to 1 inch and the number of Steps to 1000.
(e) Set Path Skip to 2 and Path Coarsen to 1.
(f) Click Display.
The spherical pathlines will be drawn along the surface.
(g) Rotate the display so that the flow field in front and in the wake of the chip
is visible, as shown in Figure 22.14.
9.00e+00
8.55e+00
8.10e+00
7.65e+00
7.20e+00
6.75e+00
6.30e+00
5.85e+00
5.40e+00
4.95e+00
4.50e+00
4.05e+00
3.60e+00
3.15e+00
2.70e+00
2.25e+00
1.80e+00
1.35e+00
9.00e-01 Y
4.50e-01 X
0.00e+00 Z

Figure 22.14: Sphere Pathlines Display
5. In the Path Lines panel, select Surface ID in the lower Color By drop-down list and
click Display (Figure 22.15).
This will color the path lines by the surface they are released from.
Note: You can also create solution animations for path lines using the Solve/Animate/
Define menu and the Animation Sequence panel.
22-26
Postprocessing
5.00e+00
4.80e+00
4.60e+00
4.40e+00
4.20e+00
4.00e+00
3.80e+00
3.60e+00
3.40e+00
3.20e+00
3.00e+00
2.80e+00
2.60e+00
2.40e+00
2.20e+00
2.00e+00
1.80e+00
1.60e+00
1.40e+00 Y
1.20e+00 X
Z
1.00e+00
Path Lines Colored by Surface ID (k)

Figure 22.15: Sphere Pathlines Colored by Surface ID

Postprocessing
Step 8: Overlaying Velocity Vectors on the Pathline Display

The overlay capability, provided in the Scene Description panel, allows you to display
multiple results on a single plot. You can exercise this capability by adding a velocity
vector display to the pathlines just plotted.
1. Enable the overlays feature.

Display −→Scene...
(a) Under Scene Composition, select Overlays.

(b) Click Apply.
2. Add a plot of vectors on the chip centerline plane.

(a) Under Options, deselect Draw Grid.
(b) Deselect all surfaces by clicking the unshaded icon to the right of Surfaces.
(c) In the Surfaces list, select fluid-sym.
(d) Set the Scale to 3.8.
Because the grid surfaces are already displayed and overlaying is active, there
is no need to redisplay the grid surfaces.
22-28
Postprocessing
(e) Click Display.

(f) Use the mouse to obtain the view that is shown in Figure 22.16.

Postprocessing
9.00e+00
8.55e+00
8.10e+00
7.65e+00
7.20e+00
6.75e+00
6.30e+00
5.85e+00
5.40e+00
4.95e+00
4.50e+00
4.05e+00
3.60e+00
3.15e+00
2.70e+00
2.25e+00
1.80e+00
1.35e+00
9.00e-01 Y
4.50e-01 X
Z
0.00e+00

Figure 22.16: Overlay of Velocity Vectors and Pathlines Display
Note: The final display (Figure 22.16) does not require mirroring about the symmetry
plane because the vectors obscure the mirrored image. You may turn off the mir-
roring option in the Views panel at any stage during this exercise.
22-30
Postprocessing
Step 9: Exploded Views

The Scene Description panel stores each display that you request and allows you to manip-
ulate the displayed items individually. This capability can be used to generate “exploded”
views, in which results are translated or rotated out of the physical domain for enhanced
display. Below, you can experiment with this capability by displaying “side-by-side” ve-
locity vectors and temperature contours on a streamwise plane in the module wake.
1. Delete the velocity vectors and pathlines from the current display.
(a) In the Names list, select the velocity vectors and pathlines.
(b) Click Delete Geometry.
(c) Click Apply.
The Scene Description panel should then contain only the two grid surfaces
(board-top and chip).
2. Create a plotting surface at x=3 inches (named x=3.0in), just downstream of the
trailing edge of the module.

Postprocessing
Hint: If you forget how to create an isosurface, see Step 3: Isosurface Creation.
3. Add the display of filled temperature contours on the x=3.0in surface.

(a) Under Options, deselect Draw Grid.
22-32
Postprocessing
(b) Deselect all surfaces by clicking on the unshaded icon to the right of Surfaces.
(c) In the Surfaces list, select x=3.0in.
(d) Click Display, and Close the panel.
The filled temperature contours will be displayed on the x=3.0 in. surface.
4. Add the velocity vectors on the x=3.0in plotting surface.

(a) Under Options, select Draw Grid and close the panel that opens.
(b) Deselect all surfaces by clicking on the unshaded icon to the right of Surfaces.
(c) In the Surfaces list, select x=3.0in.
(d) Increase the Skip to 2.
(e) Change the Scale to 1.9.
(f) Click Display.
The display will show the vectors superimposed on the contours of temperature
at x=3.0 in.

Postprocessing
5. Create the exploded view (Figure 22.17) by translating the contour display, placing
it above the vectors.
(a) In the Scene Description panel, select contour-6-temperature in the Names list.
(b) Click Transform....
This will open the Transformations panel.
(c) Under Translate, enter 1 inch for Y.

(d) Click Apply, and Close the Transformations panel.
The exploded view allows you to see the contours and vectors as distinct dis-
plays in the final scene (Figure 22.17).
6. Turn off the Overlays.

(a) In the Scene Description panel, deselect the Overlays option.
(b) Click Apply, and Close the panel.
22-34
Postprocessing
1.41e+00
1.34e+00
1.27e+00
1.20e+00
1.13e+00
1.06e+00
9.89e-01
9.20e-01
8.50e-01
7.81e-01
7.11e-01
6.41e-01
5.72e-01
5.02e-01
4.33e-01
3.63e-01
2.94e-01
2.24e-01
1.54e-01 Y
8.49e-02 X
1.53e-02 Z

Figure 22.17: Exploded Scene Display of Temperature and Velocity

Postprocessing
Step 10: Animating the Display of Results in Successive Streamwise Planes

Often, you may want to march through the flow domain, displaying a particular variable
on successive slices of the domain. While this task could be accomplished manually,
plotting each plane in turn, or using the Scene Description and Animate panels, here you
will use the Sweep Surface panel to facilitate the process. To illustrate the display of
results on successive slices of the domain, you will plot contours of velocity magnitude on
planes of constant x coordinate.
1. Delete the vectors and temperature contours from the display.

(a) In the Scene Description panel, select contour-6-temperature and vv-6-velocity-
magnitude in the Names list.
(b) Click Delete Geometry.
(c) Click Apply, and Close the panel.
The panel and display window will be updated to contain only the grid surfaces.
2. Use the mouse to un-zoom the view in the graphics window so that the entire board
surface is visible.
3. Generate contours of velocity magnitude and sweep them through the domain along
the x axis.
Display −→Sweep Surface...
(a) Keep the default Sweep Axis (the x axis).
22-36
Postprocessing
(b) Under Animation, set the Initial Value to 0 m and the Final Value to 0.1651 m.
The units for the initial and final values are in meters, regardless of the
! length units being used in the model. Here, the initial and final values are
set to the Min Value and Max Value, to generate an animation through the
entire domain.
(c) Set the number of Frames to 20.
(d) Select Contours under Display Type.
This will open the Contours panel.
i. In the Contours panel, select Velocity... and Velocity Magnitude in the

Contours Of drop-down lists.
ii. In the Contours panel, click OK.
(e) Click on Animate in the Sweep Surface panel.
You will see the velocity contour plot displayed at 20 successive streamwise planes.
FLUENT automatically interpolates the contoured data on the streamwise planes
between the specified end points. Especially on high-end graphics workstations, this
can be an effective way to study how a flow variable changes throughout the domain.
Note: You can also make use of FLUENT’s animation tools for transient cases as demon-
strated in Tutorial 4.

Postprocessing
Step 11: XY Plots

XY plotting can be used to display quantitative results of your CFD simulations. Here,
you will complete the review of the module cooling simulation by plotting the temperature
distribution along the top centerline of the module.
1. Define the line along which to plot results.

(a) In the Type drop-down list, select Line.

(b) Under End Points, enter the coordinates of the line, using a starting coordinate
of (2.0, 0.4, 0.01) and an ending coordinate of (2.75, 0.4, 0.01), as shown in
the panel above.
These coordinates define the top centerline of the module.
(c) Enter top-center-line as the New Surface Name.
(d) Click Create.
22-38
Postprocessing
2. Plot the temperature distribution along the top centerline of the module (Fig-
ure 22.18).
(a) Select Temperature... and Static Temperature in the Y Axis Function drop-down
lists.
(b) In the Surfaces list, select top-center-line.
(c) Keep the default Plot Direction of X.
This will plot temperature vs. the x coordinate along the selected line (top-
center-line).
(d) Click Axes... to modify the axis range.
This will open the Axes - Solution XY Plot panel .

Postprocessing
(e) Under Axis, select X.

(f) Under Options, deselect Auto Range.
(g) Set the Range using a Minimum of 2.0 and a Maximum of 2.75.
(h) Click Apply, and Close the panel.
(i) In the Solution XY Plot panel, click Plot.
The temperature distribution (Figure 22.18) shows the temperature increase
across the module surface as the thermal boundary layer develops in the cooling
air flow.
22-40
Postprocessing
top-center-line
4.02e+02
4.00e+02
3.98e+02
3.96e+02
Static
(k)
3.92e+02
3.90e+02
3.88e+02
1.9 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8
Y
X Position (in)
Z
Static Temperature
Figure 22.18: Temperature Along the Top Centerline of the Module

Postprocessing
Step 12: Annotation

You can annotate the display with the text of your choice.
Display −→Annotate...
1. In the Annotation Text field, enter the text describing the plot (e.g., Temperature
Along the Top Centerline).
2. Click Add.
A Working dialog box will appear telling you to select the desired location of the text
using the mouse-probe button, which is, by default, the right button.
3. Click the right mouse button in the graphics display window where you want the
text to appear, and you will see the text displayed at the desired location (Fig-
ure 22.19).
Extra: If you want to move the text to a new location on the screen, click Delete
Text in the Annotate panel, and click Add once again, defining a new position
with the mouse.
Note: Depending on the size of the graphics window and the hardcopy file format
you choose, the font size of the annotation text you see on the screen may be
different from the font size in a hardcopy file of that graphics window. The
annotation text font size is absolute, while the rest of the items in the graphics
window are scaled to the proportions of the hardcopy.
22-42
Postprocessing
top-center-line
Temperature Along the Top Centerline
4.02e+02
4.00e+02
3.98e+02
3.96e+02
Static
(k)
3.92e+02
3.90e+02
3.88e+02
1.9 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8
Y
X Position (in)
Z
Static Temperature
Figure 22.19: Temperature Along the Top Centerline of the Module

Postprocessing
Step 13: Saving Hardcopy Files

You can save hardcopy files of the graphics display in many different formats, includ-
ing PostScript, encapsulated PostScript, TIFF, PICT, and window dumps. Here, the
procedure for saving a color PostScript file is shown.
File −→Hardcopy...
1. Under Format, select PostScript.
2. Under Coloring, select Color.
3. Click Save....
This will open the Select File dialog box.
4. In the Select File dialog box, enter a name for the hardcopy file.
Summary
This tutorial has demonstrated the use of many of the extensive postprocessing features
available in FLUENT.
See Chapter 29 and Chapter 30 of the User’s Guide for more information on these and
related features.
22-44
Tutorial 23. Turbo Postprocessing
Introduction
This tutorial demonstrates the turbomachinery postprocessing capabilities of FLUENT.
In this example, you will read the case and data files (without doing the calculation) and
perform a number of turbomachinery-specific postprocessing exercises. In the process,
you will learn how to:
• Define the topology of a turbomachinery model
• Create surfaces for the display of 3D data
• Revolve 3D geometry to display a 360-degree image
• Report turbomachinery quantities
• Display averaged contours for turbomachinery
• Display 2D contours for turbomachinery
• Display averaged XY plots for turbomachinery
Prerequisites
Problem Description
The problem considered in this tutorial is a centrifugal compressor shown schematically
in Figure 23.1. The model consists of a single 3D sector of the compressor, to take
advantage of the circumferential periodicity in the problem. The flow of air through
the compressor is simulated and the postprocessing capabilities of FLUENT are used to
display realistic full 360-degree images of the solution obtained.

Turbo Postprocessing
inlet
shroud side
hub side
outlet
Setup and Solution

Preparation
1. Download turbo_postprocess.zip from the Fluent Inc. User Services Center or
2. Unzip turbo_postprocess.zip.
turbo.cas and turbo.dat can be found in the /turbo postprocess folder created
after unzipping the file.
Step 1: Reading the Case and Data Files

1. Read in the case and data files (turbo.cas and turbo.dat).
When you select turbo.cas, turbo.dat is read automatically.
23-2
Step 2: Grid Display

1. Under Options, keep the default selection of Edges.
2. Under Edge Type, select Outline.
3. Deselect all surfaces, and click on the Outline button under the Surface Types list.
4. Click Display.
5. Use the left mouse button to rotate the view, and the middle mouse button to
zoom the view to obtain an isometric display of the compressor duct, as shown in
Figure 23.2.
Extra: You can use the right mouse button to check which zone number corresponds
to each boundary. If you click the right mouse button on one of the boundaries
displayed in the graphics window, its zone number, name, and type will be printed
in the console window. This feature is especially useful when you have several zones
of the same type and you want to distinguish between them quickly.

Y
X
Z
Grid
FLUENT 6.2 (3d, coupled imp, rke)
Figure 23.2: Graphics Display of the Edges of the Compressor Mesh
23-4
Step 3: Defining the Turbomachinery Topology

In order to establish a turbomachinery-specific coordinate system, you have to define the
topology of the flow domain. The coordinate system is used in subsequent postprocessing
functions. Specifically, you will select the boundary zones that comprise the hub, shroud,
inlet, outlet, and periodics. The boundaries may consist of more than one zone.
The topology setup that you define will be saved to the case file when you save the current
model. Thus, if you read this case back into FLUENT, you do not need to set up the
topology again. See Section 27.9.1 of the User’s Guide for more information.
Define −→Turbo Topology...
1. Specify the surfaces representing the hub.

(a) Under Boundaries, keep the default selection of Hub.
(b) In the Surfaces list, select the surfaces that represent the hub (wall-diffuser-hub,
wall-hub, and wall-inlet-hub.)
2. Specify the surfaces representing the casing.

(a) Under Boundaries, select Casing.
(b) In the Surfaces list, select wall-diffuser-shroud, wall-inlet-shroud, and wall-shroud.

3. Specify the surfaces representing the periodic boundaries.

(a) Under Boundaries, select Theta Periodic.
(b) In the Surfaces list, select periodic.33, periodic.34, and periodic.35.
Note: While Theta Periodic represents periodic boundary zones on the circumfer-
ential boundaries of the flow passage, Theta Min and Theta Max are wall sur-
faces at the minimum and maximum θ position on a circumferential boundary.
There are no such wall surfaces in this problem.
4. Specify the surface representing the Inlet (inlet).
5. Specify the surface representing the Outlet (outlet).
6. Specify the surface representing the Blade (wall-blade).
7. Retain the default Turbo Topology Name of new-topology-1.
8. Click Define to set all of the turbomachinery boundaries.
FLUENT will inform you that the turbomachinery postprocessing functions have been ac-
tivated, and the Turbo menu will appear in FLUENT menu bar at the top of the console
window.
In the Turbo Topology panel, you can define any number of turbo topologies. This is
especially useful when you have a model consisting of multiple blade rows and you need to
define more than one blade row simultaneously. Each topology can be assigned a specific
name and accessed using the drop-down list in the Turbo Topology panel.
See Section 29.9.1 of the User’s Guide for more information on defining turbo topologies.
Note: You can display the selected surfaces by clicking on Display at the bottom of the
panel. This is useful as a graphical check to ensure that all relevant surfaces have
been selected.
23-6
Step 4: Isosurface Creation

To display results in a 3D model, you will need surfaces on which the data can be dis-
played. FLUENT creates surfaces for all boundary zones automatically. In a general
application, you may want to define additional surfaces for viewing results. FLUENT’s
turbo postprocessing capabilities allow you to define more complex surfaces, specific to the
application and the particular topology that you defined. In this step, you will create sur-
faces of iso-meridional (marching along the streamwise direction) and spanwise (distance
between the hub and the shroud) coordinates in the compressor.
1. Create surfaces of constant meridional coordinate.
(a) In the Surface of Constant drop-down lists, select Grid... and Meridional Coor-
dinate.
(b) Enter 0.2 under Iso-Values.
(c) Enter meridional-0.2 under New Surface Name.
(d) Click Create.
Note: The iso-values you enter for these turbo-specific surfaces are expressed
as a percentage of the entire domain (i.e., you just defined a surface of
meridional coordinate equal to 20% of the path along the duct).
(e) Repeat the steps above to define surfaces of meridional coordinates equal to
0.4, 0.6, and 0.8.

2. Create surfaces of constant spanwise coordinate.
(a) In the Surface of Constant drop-down lists, select Grid... and Spanwise Coordi-
nate
(b) Enter 0.25 under Iso-Values.
(c) Enter spanwise-0.25 under New Surface Name.
(d) Click Create.
(e) Repeat the steps above to define surfaces of spanwise coordinates equal to 0.5
and 0.75.
23-8
Step 5: Contours
1. Plot filled contours of pressure on the meridional isosurfaces (Figure 23.3).

(b) Select Pressure... and Static Pressure in the Contours Of drop-down lists.
(c) In the Surfaces list, select inlet, meridional-0.2, meridional-0.4, meridional-0.6,
meridional-0.8, and outlet.
(d) Under Options, select Draw Grid, and keep the current settings in the Grid
Display panel.
(e) Click Display in the Contours panel.
(f) Rotate and zoom the display using the left and middle mouse buttons, respec-
tively, to obtain the view shown in Figure 23.3.
In Figure 23.3, you can observe the buildup of static pressure along the duct.
2. Plot filled contours of Mach number (Figure 23.4).

(a) Select Velocity... and Mach Number in the Contours Of drop-down lists.
(b) Click Display.

1.84e+00
1.78e+00
1.73e+00
1.67e+00
1.62e+00
1.56e+00
1.50e+00
1.45e+00
1.39e+00
1.34e+00
1.28e+00
1.22e+00
1.17e+00
1.11e+00
1.06e+00
1.00e+00
9.44e-01
8.88e-01
8.32e-01
Y
7.76e-01
X
7.20e-01 Z
Contours of Static Pressure (atm)

Figure 23.3: Filled Contours of Pressure on the Meridional Isosurfaces
1.04e+00
9.85e-01
9.35e-01
8.85e-01
8.35e-01
7.84e-01
7.34e-01
6.84e-01
6.34e-01
5.83e-01
5.33e-01
4.83e-01
4.33e-01
3.82e-01
3.32e-01
2.82e-01
2.32e-01
1.81e-01
1.31e-01
Y
8.07e-02
X
3.05e-02 Z

Figure 23.4: Filled Contours of Mach Number on the Meridional Isosurfaces
23-10
In Figure 23.4, you can observe locations at which the flow becomes slightly super-
sonic, about halfway through the duct.
3. Plot filled contours of Mach number on the spanwise isosurfaces (Figure 23.5).
(a) In the Surfaces list, deselect all surfaces, and then select spanwise-0.25, spanwise-
0.5, and spanwise-0.75.
(b) Click Display.
1.04e+00
9.85e-01
9.35e-01
8.85e-01
8.35e-01
7.84e-01
7.34e-01
6.84e-01
6.34e-01
5.83e-01
5.33e-01
4.83e-01
4.33e-01
3.82e-01
3.32e-01
2.82e-01
2.32e-01
1.81e-01
1.31e-01
Y
8.07e-02
X
3.05e-02 Z

Figure 23.5: Filled Contours of Mach Number on the Spanwise Isosurfaces
The display in Figure 23.5 allows you to further study the variation of the Mach
number inside the duct. You may want to explore using different combinations of
surfaces to display the same or additional variables.
4. Display a 360-degree image of the Mach number contours on the 0.5 spanwise
isosurface (Figure 23.6).
(a) Redisplay the contours, just on the 0.5 spanwise isosurface.
i. In the Surfaces list, deselect spanwise-0.25 and spanwise-0.75.
ii. Click Display.

(b) Display the full 360-degree geometry.

i. Under Periodic Repeats, click Define... to open the Graphics Periodicity

panel.
ii. Under Cell Zones, select fluid.

This will select all the surfaces in the Associated Surfaces list. The default
value for Number of Repeats is set to 20 and the display is updated to give
full 360 degree view.
iii. Click Set in the Graphics Periodicity panel and close it.
23-12
iv. Close the Views panel.

The display will be updated to show the entire geometry.
1.04e+00
9.85e-01
9.35e-01
8.85e-01
8.35e-01
7.84e-01
7.34e-01
6.84e-01
6.34e-01
5.83e-01
5.33e-01
4.83e-01
4.33e-01
3.82e-01
3.32e-01
2.82e-01
2.32e-01
1.81e-01
1.31e-01
Y
8.07e-02 X
3.05e-02 Z

Figure 23.6: Filled Contours of Mach Number on the 0.5 Spanwise Iso Surface
Note: This step demonstrates a typical view-manipulation task. See Tutorial 22 for
further examples of postprocessing features.

Step 6: Reporting Turbo Quantities

The turbomachinery report gives you some tabulated information specific to the applica-
tion and the defined topology.
See Section 29.9.2 of the User’s Guide for details.
Turbo −→Report...
1. Under Averages, keep the default of Mass-Weighted.
23-14
2. Click Compute.
Step 7: Averaged Contours

Turbo averaged contours are generated as projections of the values of a variable averaged
in the circumferential direction and visualized on an r- z plane.
1. Turn off the periodic repeats.

(a) Under Periodic Repeats, click Define... to open the Graphics Periodicity panel.
(b) Click Reset and close the panel.
2. Display filled contours of averaged static pressure (Figure 23.7).

Turbo −→Averaged Contours...
(a) In the Contours Of drop-down lists, select Pressure... and Static Pressure.
(b) Click Display.

1.80e+00
1.76e+00
1.72e+00
1.67e+00
1.63e+00
1.58e+00
1.54e+00
1.50e+00
1.45e+00
1.41e+00
1.36e+00
1.32e+00
1.28e+00
1.23e+00
1.19e+00
1.14e+00
1.10e+00
1.06e+00
1.01e+00 Y
9.68e-01
Z X
9.24e-01
Averaged Turbo Contour - pressure (atm) (atm)

Figure 23.7: Filled Contours of Averaged Static Pressure
23-16
Step 8: 2D Contours
In postprocessing a turbomachinery solution, it is often desirable to display contours on
constant spanwise coordinates, and then project these contours onto a plane. This permits
easier evaluation of the contours, especially for surfaces that are highly three-dimensional.
FLUENT allows you to display contours in this manner using the Turbo 2D Contours panel.
1. Display 2D contours of Mach number.

Turbo −→2D Contours...
(a) In the Contours Of drop-down lists, select Velocity... and Mach Number.
(b) Under Normalised Spanwise Coordinates, enter 0.5.
Note: For highly curved edges, if a surface is created very close to the curved
edge the resulting surface may have some void spaces in it.
(c) Click Display.
(d) Use the mouse to obtain the view shown in Figure 23.8.

9.12e-01
8.69e-01
8.26e-01
7.83e-01
7.40e-01
6.96e-01
6.53e-01
6.10e-01
5.67e-01
5.24e-01
4.81e-01
4.37e-01
3.94e-01
3.51e-01
3.08e-01
2.65e-01
2.22e-01
1.79e-01
1.35e-01 X
9.22e-02 Z
Y
4.91e-02
2D Turbo Contour - mach-number

Figure 23.8: 2D Contours of Mach Number on Surface of Spanwise Value 0.5
23-18
Step 9: Averaged XY Plots

In addition to displaying data on different combinations of complex 3D and flattened
surfaces, FLUENT’s turbo postprocessing capabilities allow you to display XY plots of
averaged variables, relevant to the specific topology of a turbomachinery problem. In
particular, you will be able to plot circumferentially-averaged values of variables as a
function of either the spanwise coordinate or the meridional coordinate.
1. Plot temperature as a function of the meridional coordinate.

Turbo −→Averaged XY Plot...
(a) In the Y Axis Function drop-down lists, select Temperature... and Static Tem-
perature.
(b) In the X Axis Function drop-down list, select Meridional Distance.
(c) Under Fractional Distance, enter 0.9.
(d) Click Plot.

3.50e+02
3.40e+02
3.30e+02
3.20e+02
Static
(k)
3.00e+02
2.90e+02
2.80e+02
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Y
Z X
Meridional Distance
Averaged XY - temperature
Figure 23.9: Averaged XY Plot of Static Temperature on Spanwise Surface of 0.9 Isovalue
23-20
Summary
This tutorial has demonstrated the use of some of the turbomachinery-specific postpro-
cessing features of FLUENT. These features can be accessed once you have defined the
topology of the problem. More extensive general-purpose postprocessing features are
demonstrated in Tutorial 22.Also see Chapter 29 and Chapter 30 of the User’s Guide for
additional information.

23-22
Tutorial 24. Parallel Processing
Introduction
This tutorial illustrates the setup and solution of a simple 2D problem using FLUENT’s
parallel processing capabilities. In order to be run in parallel, the mesh must be divided
into smaller, evenly sized partitions. Each FLUENT process, called a compute node,
will solve on a single partition, and information will be passed back and forth across all
partition interfaces. FLUENT’s solver allows parallel processing on a dedicated parallel
machine, or a network of heterogeneous workstations running UNIX, or a network of
workstations running Windows.
The tutorial assumes that both FLUENT and network communication software have been
correctly installed (see the separate installation instructions and related information for
details). The case chosen is the mixing elbow problem you solved in Tutorial 1.
• Start the parallel version of FLUENT
• Partition a grid for parallel processing
• Use a parallel network of workstations
• Check the performance of the parallel solver
Prerequisites
Problem Description
The problem to be considered is shown schematically in Figure 24.1. A cold fluid at 26◦ C
enters through the large pipe and mixes with a warmer fluid at 40◦ C in the elbow. The
pipe dimensions are in inches, and the fluid properties and boundary conditions are given
in SI units. The Reynolds number at the main inlet is 2.03 × 105 , therefore, a turbulence
model will be necessary.

Parallel Processing
3
Density: ρ= 1000 kg/m
-4
Viscosity: µ = 8 x 10 Pa-s
Conductivity: k = 0.677 W/m-K 32 ″
Specific Heat: C p = 4216 J/kg-K
39.9
3°
39.9
16 ″
3°
Ux= 0.2 m/s

T = 26°C
16 ″
I = 5%
4″
12 ″ Uy= 1 m/s
32 ″ T = 40° C
I = 5%
24-2
Parallel Processing
Setup and Solution

Preparation
1. Download parallel_process.zip from the Fluent Inc. User Services Center or
2. Unzip parallel_process.zip.
elbow3.cas can be found in the /parallel process folder created after unzipping
the file.
You can partition the grid before or after you set up the problem (define models,
boundary conditions, etc.). It is best to partition after the problem is set up, since
partitioning has some model dependencies (e.g., sliding-mesh and shell-conduction
encapsulation). Since you have already followed the procedure for setting up the
mixing elbow in Tutorial 1, elbow3.cas is provided to save you the effort of re-
defining the models and boundary conditions.
Step 1: Starting the Parallel Version of FLUENT

Since the procedure for starting the parallel version of FLUENT is dependent upon the
type of machine(s) you are using, four versions of this step are provided here. Follow the
procedure for the machine configuration that is appropriate for you.
• Step 1A: Multiprocessor UNIX Machine
• Step 1B: Multiprocessor Windows Machine
• Step 1C: Network of UNIX Workstations
• Step 1D: Network of Windows Workstations

Parallel Processing
Step 1A: Multiprocessor UNIX Machine

You can start the 2D parallel version of FLUENT using 2 processes, using one of the two
methods listed below:
1. At the UNIX command prompt, type
fluent 2d -t2
2. At the command prompt, type fluent.
Do not specify any argument (e.g., 2d).

!
(a) Specify the 2D parallel version.
File −→Run...
i. Under Versions, turn on Parallel.

This is the only option that needs to be selected as the mixing elbow is a
2D problem.
ii. Under Options, specify 2 as the number of Processes.
iii. Under Options, keep the Default selection in the Communicator drop-down
list.
24-4
Parallel Processing
iv. Click Run.

Note: It is also possible to start the multiprocessor parallel version of FLUENT
from the command line instead of using the Select Solver panel.

Parallel Processing
Step 1B: Multiprocessor Windows Machine

1. At the DOS command prompt, type
fluent 2d -t2
to start the 2D parallel version with two processes.
24-6
Parallel Processing
Step 1C: Network of UNIX Workstations

1. At the command prompt, type fluent.
Do not specify any argument (e.g., 2d).

!
2. Specify the 2D network parallel version.
File −→Run...
(a) Under Versions, turn on Parallel.

This is the only option that needs to be selected as the mixing elbow is a 2D
problem.
(b) Under Options, keep the default value of 1 as the number of Processes.
You will spawn processes to other machines in the next step.
(c) Under Options, select Socket in the Communicator drop-down list.
(d) Click Run.
Note: It is also possible to start the network parallel version from the command
line instead of using the Select Solver panel.

Parallel Processing
3. Spawn one additional computational node.

Parallel −→ Network −→Configure...
(a) Specify the machine on which you want to spawn the process.
i. Under Host Entry, specify the machine name in the Hostname field.
ii. Enter your user ID in the Username field.
iii. Click Add.
The machine will be added to the Available Hosts list.
Note: It is possible to create a list of available machines and add them to the
hosts database, rather than adding machines manually. See Chapter 30 of
the User’s Guide for details.
(b) Select the newly added host in the Available Hosts list.
Note: If you do not have access to another machine to use for this tutorial,
you can spawn the second node on your own machine by selecting it from
the Available Hosts list, although you will incur a performance penalty on
a single processor machine.
(c) Under Spawn Count, keep the default value of 1.
This will give you the desired total number of 2 computational nodes.
24-8
Parallel Processing
(d) Click Spawn.

FLUENT will inform you in a Working dialog box that it is spawning the new
node. When it is done, the new node will appear in the Spawned Compute
Nodes list, as shown below.
Hint: If you accidentally spawn an undesired computational node, you can

remove it by selecting it from the Spawned Compute Nodes list and clicking
on Kill.
4. Check the network connectivity information.

Although FLUENT displays a message confirming the connection to each new com-
pute node and summarizing the host and node processes defined, you may find it
useful to review the same information at some time during your session, especially
if more compute nodes are spawned to several different machines.
Parallel −→Show Connectivity...
(a) Specify the number of the Compute Node of interest (0).

Parallel Processing
For information about all defined compute nodes, you will select node 0, since
this is the node from which all other nodes are spawned.
(b) Click Print.
--------------------------------------------------------------------
ID Comm. Hostname O.S. PID Mach ID HW ID Name
--------------------------------------------------------------------
n1 net dori hpux 11681 1 7 Fluent Node
host net bilbo hpux 12697 0 3 Fluent Host
n0* net bilbo hpux 12698 0 -1 Fluent Node
ID is the sequential denomination of each compute node (the host process is

always host), Hostname is the name of the machine hosting the compute node
(or the host process), O.S is the architecture, PID is the process ID number,
Mach ID is the compute node ID, and HW ID is an identifier specific to the
communicator used.
24-10
Parallel Processing
Step 1D: Network of Windows Workstations

The procedure below is for using the RSHD communicator software that is included with
FLUENT. You can use a different communicator if one is available on your system.
1. At the DOS command prompt, enter the following to start the 2D network parallel
version with one process:
fluent 2d -pnet -t1 -path\\machine\Fluent_Inc
where \\machine\Fluent_Inc is the path to the Fluent.Inc shared directory through

the network.
You will spawn a second compute node in the next step.
2. Spawn an additional compute node, following the procedure described in Step 1C,
substep 3, for a network of UNIX machines.
3. Check the network connectivity, following substep 4 of Step 1C.

Parallel Processing
Step 2: Reading and Partitioning the Grid

When you use the parallel solver, you need to subdivide (or partition) the grid into groups
of cells that can be solved on separate processors. If you read an unpartitioned grid into
the parallel solver, FLUENT will automatically partition it, using the default partition
settings. You can then check the partitions, to see if you need to modify the settings and
repartition the grid.
1. Inspect the automatic partitioning settings.

Parallel −→Auto Partition...
If the Case File option is turned on (the default setting), and there exists a valid
partition section in the case file (i.e., one where the number of partitions in the case
file divides evenly into the number of compute nodes), then that partition informa-
tion will be used rather than repartitioning the mesh. You need to turn off the Case
File option only if you want to change other parameters in the Auto Partition Grid
panel.
(a) Keep all defaults in the Auto Partition Grid panel and click OK.
When you keep the Case File option turned on, FLUENT will automatically
select a partitioning method for you. This is the preferred initial approach for
most problems. In the next step you will inspect the partitions created and be
able to change them, if you so choose.
2. Read the case file elbow3.cas.


24-12
Parallel Processing
Grid
Figure 24.2: Triangular Grid for the Mixing Elbow

Parallel Processing
4. Check the partition information.

Parallel −→Partition...
(a) Click Print Active Partitions.

FLUENT will print the active partition statistics to the console window.
Note: FLUENT distinguishes between two cell partition schemes within a par-
allel problem: the active cell partition, and the stored cell partition. Here,
both are set to the cell partition that was created upon reading the case
file. If you re-partition the grid using the Partition Grid panel, the new
partition will be referred to as the stored cell partition. To make it the
active cell partition, you need to click on the Use Stored Partitions button
in the Partition Grid panel. The active cell partition is used for the current
calculation, while the stored cell partition (the last partition performed) is
used when you save a case file. This distinction is made mainly to allow
you to partition a case on one machine or network of machines and solve
it on a different one.
24-14
Parallel Processing
>> 2 Active Partitions:

P Cells I-Cells Cell Ratio Faces I-Faces Face Ratio Neighbors
0 619 10 0.016 995 15 0.015 1
1 620 14 0.023 1013 15 0.015 1
----------------------------------------------------------------------
Collective Partition Statistics: Minimum Maximum Total
----------------------------------------------------------------------
Cell count 619 620 1239
Mean cell count deviation -0.1% 0.1%
Partition boundary cell count 10 14 24
Partition boundary cell count ratio 1.6% 2.3% 1.9%
Face count 995 1013 1993

Mean face count deviation -0.9% 0.9%
Partition boundary face count 15 15 15
Partition boundary face count ratio 1.5% 1.5% 0.8%
Partition neighbor count 1 1

----------------------------------------------------------------------
Partition Method Principal Axes
Stored Partition Count 2
Done.
(b) Review the partition statistics.

An optimal partition should produce an equal number of cells in each parti-
tion for load balancing, a minimum number of partition interfaces to reduce
interpartition communication bandwidth, and a minimum number of partition
neighbors to reduce the startup time for communication. Here, you will be
looking for relatively small values of mean cell and face count deviation and
total partition boundary cell and face count ratio.

Parallel Processing
5. Examine the partitions graphically.

(a) Initialize the solution.
In order to use the Contours panel to inspect the partition you just created,
you have to initialize, even though you are not going to solve the problem at
this point. For this initialization, default values are sufficient.
(b) Display the cell partitions (Figure 24.3).
i. In the Contours Of drop-down lists, select Cell Info... and Active Cell
Partition.
ii. Under Options, select Filled.
iii. Set the number of Levels to 2, the number of compute nodes.
iv. Click Display.
As shown in Figure 24.3, the cell partitions are acceptable for this problem.
The position of the interface reveals that the criteria mentioned above will be
matched. If you were unsatisfied with the partitions, you could use the Partition
Grid panel to repartition the grid. Recall that, if you wish to use the modified
partitions for a calculation, you will need to make the Stored Cell Partition the
Active Cell Partition by either clicking on the Use Stored Partitions button in the
Partition Grid panel or saving the case file and reading it back into FLUENT.
24-16
Parallel Processing
1.00e+00
5.00e-01
0.00e+00
Contours of Active Cell Partition

Figure 24.3: Cell Partitions
See Section 32.5.3 of the User’s Guide for details about the procedure and
options for manually partitioning a grid.
6. Save the case file with the partitioned mesh (elbow4.cas).


Parallel Processing
Step 3: Solution

(b) Click on Init and Close the panel.
A warning message will appear asking you to discard the data that was gen-
erated during the first initialization, which was used to inspect the Contours
panel. Accept to discard the data.


The solution will converge in approximately 80 iterations.
4. Save the data file (elbow4.dat).

24-18
Parallel Processing
Step 4: Checking Parallel Performance

Generally, you will use the parallel solver for large, computationally intensive problems,
and you will want to check the parallel performance to determine if any optimization is
required. Although the example in this tutorial is a simple 2D case, here you will check
the parallel performance as an exercise.
Parallel −→ Timer −→Usage
Performance Timer for 86 iterations on 2 compute nodes

Average wall-clock time per iteration: 0.023 sec
Global reductions per iteration: 91 ops
Global reductions time per iteration: 0.000 sec (0.0%)
Message count per iteration: 69 messages
Data transfer per iteration: 0.006 MB
LE solves per iteration: 6 solves
LE wall-clock time per iteration: 0.006 sec (25.7%)
LE global solves per iteration: 2 solves
LE global wall-clock time per iteration: 0.000 sec (0.6%)
AMG cycles per iteration: 11 cycles
Relaxation sweeps per iteration: 303 sweeps
Relaxation exchanges per iteration: 86 exchanges
Total wall-clock time: 1.966 sec

Total CPU time: 3.570 sec
The most accurate way to evaluate parallel performance is by running the same paral-
lel problem on 1 CPU and on n CPUs, and comparing the Total wall-clock time
(elapsed time for the iterations) in both cases. Ideally you would want to have the Total
wall-clock time with n CPUs be 1/n times the Total wall-clock time with 1 CPU.
In practice, this improvement will be reduced by the performance of the communication
subsystem of your hardware, and the overhead of the parallel process itself. As a rough
estimate of parallel performance, you can compare the Total wall-clock time with
the CPU time. In this case, the CPU time was approximately 1.98 times the Total
wall-clock time. For a parallel process run on two compute nodes, this reveals very
good parallel performance, even though the advantage over a serial calculation is small,
as expected for this simple 2D problem.

Parallel Processing
See Tutorial 1 for complete postprocessing exercises for this example. Here, two plots are
generated so that you can confirm that the results you obtained with the parallel solver
are the same as those you obtained with the serial solver.
1. Display an XY plot of temperature across the exit (Figure 24.4).

Plot −→ XY Plot...
(a) Select Temperature... and Static Temperature in the Y Axis Function drop-down
lists.
(b) Select pressure-outlet-7 in the Surfaces list.
(c) Click on Plot.
24-20
Parallel Processing
pressure-outlet-7
3.12e+02
3.10e+02
3.08e+02
3.05e+02
Static 3.02e+02
Temperature
(k) 3.00e+02
2.98e+02
2.95e+02
2.92e+02
48 50 52 54 56 58 60 62 64
Position (in)
Static Temperature
Figure 24.4: Temperature Distribution at the Outlet

Parallel Processing
2. Display filled contours of the custom field function dynam-head (Figure 24.5).
(a) Select Custom Field Functions... in the drop-down list under Contours Of.
The function you created in Tutorial 1, dynam-head, will be shown in the lower
drop-down list.
(b) Change the number of Levels back to 20.
(c) Click on Display, and then Close the panel.
24-22
Parallel Processing
8.06e+02
7.65e+02
7.25e+02
6.85e+02
6.44e+02
6.04e+02
5.64e+02
5.24e+02
4.83e+02
4.43e+02
4.03e+02
3.62e+02
3.22e+02
2.82e+02
2.42e+02
2.01e+02
1.61e+02
1.21e+02
8.06e+01
4.03e+01
0.00e+00
Contours of dynam-head
Figure 24.5: Contours of the Custom Field Function, Dynamic Head
Summary
In this tutorial you learned how to solve a simple 2D problem using FLUENT’s parallel
solver. Here the automatic grid partitioning performed by FLUENT when you read the
mesh into the parallel version was found to be acceptable. You also learned how to check
the performance of the parallel solver to determine if optimizations are required.
See Section 32.6 of the User’s Guide for additional details about using the parallel solver.

Parallel Processing
24-24

Fluent Problems

Transféré par

Informations du document

Description originale:

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Fluent Problems

Transféré par

Droits d'auteur :

Formats disponibles

FLUENT 6.

CHEMKIN is a registered trademark of Reaction Design Inc.

Portions of this program include material copyrighted by PathScale Corporation

What’s In This Manual

Where to Find the Files Used in the Tutorials

How To Use This Manual

For the Beginner

For the Experienced User

Typographical Conventions Used In This Manual

• An informational icon ( i ) marks an important note.

• An warning icon ( ! ) marks a warning.

1 Introduction to Using FLUENT: Fluid Flow and Heat Transfer

2 Modeling Periodic Flow and Heat Transfer 2-1

Step 3: Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-8

3 Modeling External Compressible Flow 3-1

4 Modeling Unsteady Compressible Flow 4-1

Step 4: Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-8

5 Modeling Radiation and Natural Convection 5-1

Step 13: Comparison of Radiation Models for a Non-Participating

6 Using a Non-Conformal Mesh 6-1

7 Using a Single Rotating Reference Frame 7-1

Step 2: Units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-5

8 Using Multiple Rotating Reference Frames 8-1

9 Using the Mixing Plane Model 9-1

10 Using Sliding Meshes 10-1

Step 8: Solution: Steady Flow with Non-Moving Rotor . . . . . . . . . . 10-19

11 Using Dynamic Meshes 11-1

12 Modeling Species Transport and Gaseous Combustion 12-1

Setup and Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-3

13 Using the Non-Premixed Combustion Model 13-1

14 Modeling Surface Chemistry 14-1

15 Modeling Evaporating Liquid Spray 15-1

Step 5: Create a Spray Injection . . . . . . . . . . . . . . . . . . . . . . . 15-24

16 Using the VOF Model 16-1

17 Modeling Cavitation 17-1

Step 4: Phases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-11

18 Using the Mixture and Eulerian Multiphase Models 18-1

19 Using the Eulerian Multiphase Model for Granular Flow 19-1

Step 1: Grid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-3

20 Modeling Solidification 20-1

Step 2: Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-5

23 Turbo Postprocessing 23-1

24 Parallel Processing 24-1

Step 3: Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24-18

• Read an existing grid file into FLUENT

• Use mixed units to define the geometry and fluid properties

• Initiate the calculation with residual plotting

• Calculate a solution using the segregated solver

• Examine the flow and temperature fields using graphics

• Enable the second-order discretization scheme for improved prediction of tempera-

Figure 1.1: Problem Specification

Setup and Solution

where cdrom must be replaced by the name of your CD-ROM drive.

3. Start the 2D version of FLUENT.

2. Check the grid.

3. Smooth (and swap) the grid.

4. Scale the grid.

5. Display the grid (Figure 1.2).

Figure 1.2: The Triangular Grid for the Mixing Elbow

2. Turn on the standard k- turbulence model.

3. Enable heat transfer by activating the energy equation.

(c) Click Change/Create.

(d) Click No when FLUENT asks if you want to overwrite air.

2. Turn on the standard k- turbulence model.