Professional Documents
Culture Documents
Theory Manual
Version 5
October 1998
CFDRC Confidential
Table of Contents
List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
List of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi
Nomenclature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii
Roman Symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii
Greek Symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv
Subscripts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvii
Superscripts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvii
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xix
Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xix
Audience . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xx
Summary of Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .xxi
Grids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxi
Phenomena Simulated . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxi
Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxii
Density Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxii
Viscosity Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxii
Specific Heat Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxii
Fluid Conductivity Options. . . . . . . . . . . . . . . . . . . . . . . . . . . xxii
Species Mass Diffusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxiii
Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxiii
Turbulence Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxiii
Gas Phase Chemical Reaction Models . . . . . . . . . . . . . . . . . xxiii
Surface Reaction Models . . . . . . . . . . . . . . . . . . . . . . . . . . . xxiii
Spray Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxiii
iii
Table of Contents
Domain Modeling
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Structured Grids. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Multi-Domain Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Coordinate Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Cartesian Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Body-Fitted Coordinates. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Blockage Concept . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Chapter 2
Chapter 3
1-1
1-1
1-2
1-4
1-4
1-4
1-8
2-1
2-1
2-2
2-3
2-4
2-5
2-9
Discretization Methods
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-1
Finite Volume Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-1
Transient Term. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-3
Convection Term . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-3
First-Order Upwind Scheme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-5
Central Difference Scheme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-6
Second-Order Upwind Scheme. . . . . . . . . . . . . . . . . . . . . . . . . . . 3-7
Smart Scheme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-7
Other Higher-Order Schemes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-8
Osher-Chakravarthy Scheme . . . . . . . . . . . . . . . . . . . . . . . . . 3-8
Diffusion Terms. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-8
Source Term . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-10
iv
Table of Contents
Velocity-pressure Coupling
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Continuity and Mass Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Pressure -correction Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
SIMPLEC Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Chapter 5
Special Topics
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Moving Grid Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Conjugate Heat Transfer Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Interface Constraints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Equivalent Thermal Conductivity. . . . . . . . . . . . . . . . . . . . . . . . .
Turbulent Flow Considerations. . . . . . . . . . . . . . . . . . . . . . . . . . .
Chapter 6
5-1
5-1
5-4
5-5
5-6
5-8
Boundary Conditions
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
General Treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Inlet Boundaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Prescribed Mass Flux Inlet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Prescribed Total Pressure Inlet . . . . . . . . . . . . . . . . . . . . . . . . . . .
Exit Boundaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Prescribed Pressure Exit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Extrapolated Conditions Exit . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Combined Fixed Pressure and Extrapolated Exit . . . . . . . . . . . . .
Symmetry Boundaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Wall Boundaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Momentum Equations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Enthalpy Equation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Turbulence Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Chapter 7
4-1
4-1
4-6
4-6
6-1
6-1
6-2
6-2
6-3
6-5
6-5
6-6
6-6
6-7
6-7
6-7
6-8
6-9
Solution Methods
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Overall Solution Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
SIMPLEC Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
PISO Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Under-relaxation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7-1
7-1
7-1
7-4
7-6
Table of Contents
Turbulence
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-1
Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-1
Baldwin-Lomax Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-3
Standard k-e Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-5
RNG k- Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-9
Kato-Launder k-e Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-10
k- Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-11
Low Reynolds Number k- Model . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-12
Chapter 9
Reaction Models
Basic Definitions And Relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-1
Composition variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-2
Chemical rate expressions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-3
Mixture fractions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-6
Instantaneous Chemistry Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-8
Single-Step Finite Rate Chemistry Model. . . . . . . . . . . . . . . . . . . . . . 9-10
Equilibrium Chemistry Model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-11
Multi-Step Finite-Rate Reaction Model . . . . . . . . . . . . . . . . . . . . . . . 9-14
Surface Reaction Models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-15
Turbulence-combustion Interaction. . . . . . . . . . . . . . . . . . . . . . . . . . . 9-16
Determination of PDF. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-17
Reaction Progress PDF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-18
Mixture Fraction PDF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-18
Determination of Averaged Variables. . . . . . . . . . . . . . . . . . . . . 9-21
Chapter 10
Spray Model
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Size Distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Droplet Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Droplet Evaporation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Droplet Source/Sink Terms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Mass Conservation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Momentum Conservation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
vi
10-1
10-2
10-4
10-6
10-8
10-9
10-9
Table of Contents
Energy Conservation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Droplet Dispersion Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Droplet Break-up Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Droplet Collision/Coalescence Model. . . . . . . . . . . . . . . . . . . . . . . .
Chapter 11
10-10
10-11
10-12
10-16
Radiation Model
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-1
Blackbody Radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-1
Radiation Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-4
Radiation Characteristics of Gases . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-6
Radiative Transfer Equation (RTE). . . . . . . . . . . . . . . . . . . . . . . . . . . 11-6
Discrete Ordinate Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-9
Conjugate Heat Transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-12
AAQ Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-13
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References-1
Appendix A
Fluid Properties
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-1
Density. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-1
Constant Density. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-1
Perfect Gas Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-1
Inverse Temperature Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-2
Viscosity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-2
Constant Viscosity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-2
Power Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-2
Sutherlands Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-2
Specific Heat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-3
Constant Specific Heat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-3
Thermal Diffusivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-3
Mass Diffusivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-3
Appendix B
JANNAF Tables
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B-1
Appendix C
vii
Table of Contents
viii
List of Figures
Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xix
Figure 1. Schematic Representation of CFD-ACE . . . . . . . . . . . . . . . . . . xx
Figure 2. Three-Dimensional Control Volume. . . . . . . . . . . . . . . . . . . . .xxvi
Chapter 1
Domain Modeling
Figure 1-1.
Figure 1-2.
Figure 1-3.
Figure 1-5.
Chapter 3
1-2
1-3
1-4
1-7
Discretization Methods
Figure 3-1. Schematic: A control-volume in 3-D body-fitted Coordinate . 3-2
Figure 3-2. 2-D Stencil: Discretization of Convection and Diffusion Terms 3-5
Chapter 5
Special Topics
Figure 5-1. Schematic: Flow Field Around Automobile During Defrost . . 5-5
Figure 5-2. Interfacial Conditions for Conjugate Heat Transfer . . . . . . . . . 5-6
Chapter 6
Boundary Conditions
Figure 6-1. Boundary Cell. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-2
Chapter 7
Solution Methods
Solution Flowchart for SIMPLEC Algorithm . . . . . . . . . . . . . . . . . . . . . . . 7-3
Figure 7-2. Solution Procedure for a Single Dependent Variable . . . . . . . . 7-4
ix
List of Figures
Radiation Model
Figure 11-1. Spectral Emissive Power of a Blackbody at Different
Temperatures (Siegel and Howell) . . . . . . . . . . . . . . . . . . . . . 11-3
Figure 11-2. Schematic Depicting the Radiative Heat Transfer in an EmittingAbsorbing and Scattering Gray Medium . . . . . . . . . . . . . . . . 11-8
List of Tables
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xix
Table 1. Boundary Conditions in CFD-ACE Structured Solver . . . . . . . . xxiv
Chapter 11
Radiation Model
Table 11-1. Direction Cosines and Their
Associated Weights for the S4 Approximation . . . . . . . . . . . .11-11
xi
List of Tables
xii
Nomenclature
Roman Symbols
area
link coefficient
CP
C1
C2
ei
ei
xiii
Nomenclature
xiv
Roman Symbols
mixture fraction
radiation intensity
thermal conductivity
molecular weight
mass
ni
pressure
heat flux
Re
Reynolds number
droplet radius
temperature
T+
Greek Symbols
Greek Symbols
Nomenclature
time
ui
friction velocity
u+
velocity vector
We
Weber number
Cartesian coordinate
xi
Yi
Cartesian coordinate
y+
Cartesian coordinate
direction cosine
xv
Nomenclature
xvi
Greek Symbols
diffusion coefficient
direction cosine
curvilinear coordinate
curvilinear coordinate
element potential
curvilinear coordinate
density
ij
solid angle
frequency
CFD-ACE Theory Manual
Subscripts
Subscripts
Superscripts
Nomenclature
curvilinear coordinate
black body
droplet
time level
wall units
xvii
Nomenclature
xviii
Superscripts
Introduction
Overview
xix
Introduction
Audience
CFD-GUI
Graphical User
Interface
Input File
Preprocessor
Structured
Multi-Domain
Grid
Graphics
Files
Results
File
Analysis
Audience
xx
Summary of Features
Summary of
Features
Introduction
Grids
Grids
Phenomena Simulated
Properties
Models
Boundary Conditions
Solution Methodology
CFD-ACE employs a structured, multi-domain, body-fitted-coordinated-system approach to enable the user to simulate flows in complex geometric configurations. Solid obstacles to flow are easily
modeled with blockage concept.
While the CFD-ACE structured flow solver is very versatile with
respect to the geometric complexity it can accommodate, there is little
geometric modeling and grid generation capability built into the program. CFD-ACE assumes that suitable grids have been constructed
elsewhere and can be imported. The primary grid generator supported
by this program is CFDRCs CFD-GEOM [4]. However, CFD-ACE
structured flow solver can read grids in Plot3D format, and it also has
interfaces to the ICEM-CFD/CAE and GRIDGEN programs.
Phenomena
Simulated
xxi
Introduction
Summary of Features
can be incorporated to provide conjugate heat transfer capability. Radiative heat transfer can be accounted for as well.
Mixing and Reacting Flows. This type of problem requires the
solution of additional equations for mixture fractions or species
mass fractions.
Steady-state or Transient. For all types of problems, phenomena can be simulated in either steady-state or time-dependent
fashion.
Incompressible or Compressible. Flow speeds for problems can
range from low speed subsonic to transonic and supersonic.
Properties
Several options for fluid density, viscosity, specific heat, fluid conductivity, and species mass diffusion are available within the CFDACE structured flow solver.
Density Options
Constant
Temperature-dependent
Pressure-dependent
Ideal Gas Law
Constant
Sutherland's Law (direct input of constants)
Sutherland's Law (constants determined from local mixture)
Polynomial in Temperature
Kinetic Theory of Gases (for multicomponent mixtures)
Non-Newtonian Power Law
Non-Newtonian Carreau Law
Constant
Calculation from JANNAF relations based on local mixture
Fluid Conductivity
Options
Constant
Constant Prandtl Number
Polynomial in Temperature
xxii
Summary of Features
Introduction
Models
Turbulence Models
Instantaneous
Equilibrium
One-step Finite Rate
Multi-step Finite Rate
Spray Model
xxiii
Introduction
Summary of Features
Radiation Models
Boundary
Conditions
Variations
Inlets
Outlets
Walls
Thin Walls
--
Source Terms
xxiv
Symmetry
--
Mixing plane
--
Periodic
Summary of Features
Solution
Methodology
Introduction
Control Volume
Approach
xxv
Introduction
Summary of Features
L
P
S
Figure 2. Three-Dimensional Control Volume
Pressure-Based Solution
Algorithm
Iterative Scheme
CFD-ACE employs an iterative solution scheme in which the assembled equations for each dependent variable are solved sequentially
and repeatedly with the goal of improving the solution at each iteration, finally reducing errors to acceptably small values. The nonlinear
and coupled nature of the equations makes it necessary to restrain or
under-relax the iteration-to-iteration changes in the variables in order
to obtain convergence of the solution procedure.
xxvi
Introduction
Domain Modeling
Structured Grids
The CFD-ACE structured flow solver supports structured grid topologies. Structured grids use a structured arrangement of control volumes, or cells, to discretize the problem domain. The cells are
arranged into rows, columns, and layers. The position of each cell,
relative to other cells, is uniquely determined by a series of three integers (I, J, K), corresponding to the row, column, and layer in which
the cell resides.
A single structured grid consists of a structured arrangement of cells
which form a computational square for two-dimensional models and a
computational cube for three-dimensional models. The two-dimensional model is made up of L x M control cells where L is the number
1-1
Multidomain Approach
of cells in the I-direction and M is the number of cells in the J-direction. The three-dimensional model consists of L x M x N control cells
where L and M are defined as for the two-dimensional model, and N
is the number of cells in the K-direction. Two-dimensional and threedimensional structured grids are shown in figure 1-1. Note that in
each direction, there is one more cell vertex than cells.
J =M
K=N
K= 1
J=1
I=1
I=L
J =M
I=L
J=1 I=1
Multidomain
Approach
1-2
Construction of a single structured grid may be a difficult task in complex configurations. For illustrative purposes, consider the flow over a
backward-facing step. If a single structured grid is constructed, it
would be highly skewed behind the step as can be seen in figure 1-2a.
To circumvent this difficulty, CFD-ACE employs a powerful methodology known as the multidomain approach. If the flow domain for the
Multidomain Approach
backstep problem is divided into two domains, grids can be conveniently constructed in each domain as shown in figure 1-2b.
The grids must not overlap and must be joined at their boundary
surfaces.
The boundary surfaces of each cell in grids on opposite sides of a
boundary must match exactly.
1-3
Coordinate Systems
Coordinate
Systems
In CFD-ACE, two domain coordinate systems are available: Cartesian or body-fitted (BFC).
Cartesian
Coordinates
In the Cartesian system, the I grid direction is aligned with the X-axis,
the J grid direction is aligned with the Y-axis, and the K grid direction
is aligned with the Z-axis. (Two-dimensional Cartesian grids are thus
restricted to the X-Y plane.)
Body-Fitted
Coordinates
A body-fitted grid is one in which the I-, J-, and K-direction grid lines
are allowed to bend and curve through space to conform to the boundaries of the flow domain (see figure 1-3). (Two-dimensional body-fitted grids are also restricted to the X-Y plane.)
In general in BFC systems, grid lines in the same direction are not
parallel and grid lines in different directions are not perpendicular. For
illustrative purposes, let us consider a 2-D BFC grid and the corre-
1-4
Coordinate Systems
M+1
cell, (I,J)
J,
3
I,,
2
J=1
I= 1
x
0
(a) Physical Domain
2...
...L L+1
x = x ( , , )
= ( x, y, z ), or y = y ( , , )
= ( x, y, z ), z = z ( , , )
(1-1)
= ( x, y, z ),
A few basic relations between the Cartesian and BFC systems will be
presented below. Interested users should refer to Thompson et al,
CFD-ACE Theory Manual
1-5
Coordinate Systems
(1985) for a comprehensive mathematical treatment of the BFC system. To facilitate the use of tensor notation, (1, 2, 3) will be used to
denote (, , ). In a BFC coordinate system, the covariant base vectors are defined as
r
j = -------- ; j = ( 1, 2, 3 )
j
(1-2)
where r is the position vector and is equal to xi + yj + zk, and contravariant base vectors are defined as
2 3 2 3 1 3 1 2
1
= ---------------; = ---------------- ; = ---------------J
J
J
(1-3)
1-6
Coordinate Systems
i = ij
(1-4)
where ij is the Kronecker delta. But note that j and j are not necessarily unit vectors. Covariant base unit vectors can be defined as
j
e j = ---- , h j = j =
hj
x 2 y 2 z 2
-------+ -------- + ------- 1
2
3
(1-5)
1-7
Blockage Concept
1
k
f = --- -------- ( J f )
J k
Gradient:
Divergence:
Curl:
1
k
V = --- -------- ( J V )
J k
1
k
V = --- -------- ( J V )
J k
(1-6)
(1-7)
(1-8)
Laplacian:
Blockage Concept
1-8
1
i
j
2f = --- ---------------- [ Jf ]
J i j
(1-9)
Blockage Concept
Figure 1-6. Illustration of Blockage Concept for the Flow over a Cylinder in a Channel
1-9
1-10
Blockage Concept
Introduction
In the CFD-ACE methodology, fluid flows are simulated by numerically solving partial differential equations that govern the transport of
flow quantities also known as flow variables. These variables include
mass, momentum, energy, turbulence quantities, mixture fractions,
species concentrations, and radiative heat fluxes. The variables for
which transport equations have to be solved will depend on the nature
of the flow problem. In this chapter, the basic governing equations
for the conservation of mass, momentum and energy are presented.
Chapters 8 and 9 introduce the transport equations for other flow variables.
CFD-ACE employs conservative finite-volume methodology and
accordingly all the governing equations are expressed in conservative
form. Cartesian coordinate system and tensor notation are generally
employed in which repeated indices imply summation over all coordinate directions. The user should note that CFD-ACE solves the governing equations in a cylindrical coordinate system for 2-D axis
symmetric flow problems. These equations can be found in any fluid
mechanics text (e.g., Wylie, 1974 and Schlichting, 1979) and therefore are not presented here.
Continuity
Equation
In any fluid flow in which the continuum hypothesis holds mass conservation can be expressed as
2-1
Continuity Equation
----------- + - ( u j ) = m
x
t
j
(2-1)
where uj is the j th Cartesian component of the instantaneous velocity, is the fluid density, and m represents the rate of mass generation in the system. Typical examples of mass generation include
surface reaction, spray vaporization etc.
Momentum
Equations
p ij
---- ( u i ) + ------- ( u i u j ) = ------ + --------- + f i
t
x i
x i x i
(2-2)
u u 2 u k
ij = -------i + -------j --- -------- ij
x j x i 3 x k
(2-3)
p
---- ( u i ) + ------- ( u i u j ) = -----t
x j
x i
u u 2 k (2-4)
+ ------- -------i + -------j --- --------- ij
x j x i x i 3 x k
+ f i
2-2
Continuity Equation
Energy Equation
The equation for the conservation of energy can take several forms.
In CFD-ACE, the user may choose either static enthalpy or stagnation enthalpy depending on the application. The static enthalpy form
of the energy equation can be expressed as
u i
q p
p
( h ) + ------- ( u j h ) = -------j + ------ + u j ------- + ij ------x j
x j
t
x j
x j t
------- ( J ij h i ) + S a
x j
(2-5)
Here, J ij is the total (concentration-driven + temperature-driven) diffusive mass flux for species i which will be given later, h i represents
the enthalpy for species i, S a stands for additional sources due to
surface reaction, radiation and liquid spray and, qj is the j-component of the heat flux. Fouriers Law is employed to model the heat
flux
T
q j = K ------x j
(2-6)
uj u j
H = h + -------2
(2-7)
2-3
Continuity Equation
T
---- ( H ) + ------- ( u j H ) = ------- K ------- + p
------ + ------- ( u )
t
x j
x j x j t x i ij j
------- ( J ij h i ) +S a + f i u i
x j
Rotating Frame of
Reference
(2-8)
vi = ui + wi
Since the entrainment velocity does not contribute to the mass balance, the continuity equation, equation 2-1, remains invariant.
Regarding the momentum equation, equation 2-2, it is well known
that observers in the two systems of reference will not see the same
field of forces since the inertia term, i.e. the RHS of equation 2-2,
changes when passing from one system to the other. It can be shown
that under a steadily rotating system the momentum equation for the
relative velocity components goes as follows,
p ij
---- ( u i ) + ------- ( u i u j ) = ------ + --------- + f i 2 ipj p u j
x i
t
x i x i
(2-9)
( j x j i j x i j )
here ipj is the alternating tensor. Compared with equation 2-2, the
above equation has two additional terms on the right hand side, the
2-4
Continuity Equation
first one is known as the Coriolis force and the second one is referred
to as the centrifugal force. These two forces, acting on a fluid particle
the rotating system, play a very important role in rotating flows.
With the understanding that u stands for the relative velocity and
t is with reference to the rotating frame too equation 2-5 is still
valid for static enthalpy.
However, the total enthalpy equation, equation 2-8, is no longer valid
for the rotating system. In order to arrive at a new total enthalpy
equation, we simply repeat the process in which equation 2-8 has
been obtained. First multiply the momentum equation (2-9) by u i and
then add to the static enthalpy equation (2-5) we have,
p
---( I ) + ------- ( u j I ) = ------- K ------- + ------ + ------- ( ij u j )
t
x x j t x i
x j
j
------- ( J ij h i ) + S a + f i u i
x j
(2-10)
ui ui wi wi
I h + -------- ----------- = H w i v i
2
2
(2-11)
Favre Averaged
Equations
The fundamental equations of fluid dynamics that have been introduced in the preceding sections are, in general, applicable to Newtonian fluid flow under steady or transient, incompressible or
compressible, laminar, transitional or turbulent conditions. The nonlinearity of the Navier-Stokes equations, coupled with the complexity
of the boundary conditions, makes it impossible to obtain analytical
solutions for all but a limited number of flows of engineering interest.
2-5
Continuity Equation
1
= + where = --- ( t o + T ) dt
T
to
(2-12)
= + where =
(2-13)
2-6
Continuity Equation
----------- + - ( u j ) = m
t x j
(2-14)
Similarly when the Navier-Stokes equations equation 2-2 are averaged, we obtain the Favre-averaged Navier-Stokes (FANS) equations
given below. (For detailed derivation, see Cebeci and Smith, 1974.)
p
---- ( u i ) + ------- ( ui uj ) = -----t
xj
xi
u u 2 u m
+ ------- u i u j
x j
The FANS equations contain less information than the full NS equa-
u u 2 u m 2
ui uj = t -------i + -------j --- ---------- ij --- k ij
x j x i 3 x m 3
(2-16)
2-7
Continuity Equation
1
k = --- u ku k
2
(2-17)
By substituting equation 2-16 into equation 2-15, we obtain the modeled FANS equations
p
---- ( u j ) + ------- u i u j = -----t
x j
x i
u u 2 u m
p
T
h -----p
--------- + - ( u j h ) = ------- K ------- + ------ + u j ------x j x j t
x j
x j
t
ui
+ ij ------- ------- ( J ij h i ) + S a
x j x j
(2-19)
u i
------- uj h + ij ---------x j
x j
h
T
uj h = t ------- = C p t -----xj
xj
2-8
(2-20)
Continuity Equation
t = -----t
t
The second term arises from the viscous dissipation and is modeled as
follows:
When the rate of dissipation of turbulent kinetic energy, , is solved
u i
ij ---------- =
x j
u i
u i u u 2 u k
ij ---------- = t ------- -------i + -------j --- -------- ij
x j
x j x j x i 3 x k
T p
H
) = ------ K ------- + ------ + ------- ( ij uj )
----------- + ------- ( u j H
x j x j
x j
t x i
t
------- ( J ij h i ) + S a + fi ui
x j
(2-21)
Generalized
Transport Equation
The user may have noted that all governing equations, except the continuity equation, possess a common form and can be expressed in a
general form as
2-9
Continuity Equation
---- ( ) + ------- ( u j ) = ------- ------- + S
x j
t
x j x j
Transient
Convection
Diffusion
(2-22)
Source
The equations for the other variables such as the turbulence quantities, species mass fractions which will be introduced in chapters 9 and
10 also exhibit the same form. Thus, the symbol may represent any
of the velocity components, enthalpy, or other scalar flow variables.
In the preceding equation, the overbar and tilde have been dropped for
convenience.
In equation 2-23, is the effective diffusivity and is modeled as
= --- + -----t
t
(2-23)
2-10
Discretization Methods
Introduction
Finite Volume
Approach
1
1
1
k
ik
--- ---- ( J )+--- -------- { J ( V ) } = --- -------- Jg ------- + S (3-1)
J k
J t
J k
j
CFD-ACE Theory Manual
3-1
with
jk
(3-2)
The user should note that tensor notation is used for the sake of brevity. Throughout this chapter, (, , ) and (1, 2, 3) may be used
interchangeably.
It has been noted in Chapter 1 that the flow domain is divided into a
number of cells known as control volumes. In the finite volume
approach of CFD-ACE, the governing equations are numerically integrated over each of these computational cells or control volumes. Let
us consider one such control volume as shown in figure 3-1 .
N
L
E
P
W
3-2
Transient Term
variables and fluid properties are stored at the cell center P. In other
words, the average value of any flow quantity within a control volume
is given by its value at the cell center. The finite-volume numerical
method essentially involves the integration of equation 3-1 over this
control volume.
[Equation3-1] Jd d d
(3-3)
Transient Term
--- ---( J ) Jd d d =
J t
t
( J ) d d d
(3-4)
o o
= ------------------------------------t
Convection Term
3-3
Convection Term
1
k
C = --- -------- [ J ( V ) ]
J k
1 J
1
1 J
2
1 J
3
= --- ------ ----- U + --- ------ ----- U + --- ------ ----- U
J h 1
J h 2
J h 3
(3-5)
Ce Cw + Cn Cs + Ch Cl = G e e G w w + G n n
(3-6)
Gss + Ghh Gll
with Ge, for example, defined as
J
1
G e = ----- U
h
1
e
(3-7)
3-4
Convection Term
N
NE
NW
n
e
w
SW
SE
S
First-Order Upwind
Scheme
UP
Ce
G p if G e > 0
= e
G e E if G e < 0
(3-8)
Ce
E + p
E p
= G e ------------------ G e -----------------2
2
(3-9)
3-5
Convection Term
As its name implies, this scheme has first-order accuracy and is one of
the most stable schemes.
Central Difference
Scheme
Ce
E + p
= G e -----------------2
(3-10)
C e = de C e
CD
+ ( 1 d e )C e
(3-11)
p + E
E p
C e = G e ------------------ d e G e -----------------2
2
(3-12)
3-6
Convection Term
Second-Order
Upwind Scheme
In this scheme the cell face value is evaluated using two upstream
nodes. The cell face convective flux is given by
Ce =
3
1
G e --- E --- EE if G e < 0
2
2
Smart Scheme
(3-13)
(3-14)
d e = 1 Minimode ( 1, r e )
(3-15)
3-7
Diffusion Terms
i + 1 u i u
r e = ------------------------------------- with u = sign ( G e )
i + 1 i
(3-16)
Other Higher-Order Presently one higher order scheme with limiters is available in CFDACE for accurate shock capturing in high-speed flows. This scheme
Schemes
can be represented by equation 3-12 but with different expressions
for de.
Osher-Chakravarthy
Scheme
1
d e = 1 ------------ Minimode ( , r e )
2
1+
+ ------------- Minimode ( 1, r e )
2
1
3
(3-17)
3
1
Diffusion Terms
The diffusion term in equation 3-1 can be split into two parts: main
diffusion (j = k); and cross diffusion (j k). Lets consider main diffusion first, i.e.,
1
k
kk
D m = --- -------- Jg -------- , k = 1, 2, or 3
J k
k
(3-18)
3-8
Diffusion Terms
D m J d
11
11
d d = Jg ------ Jg ----- e
w
(3-19)
Jg
11
A
= -------------------h 1 sin 1
(3-20)
A
D = --------------------------h 1 sin 1
(3-21)
we have
1
Dm J d d d
= De ( E p ) D w ( p W )
(3-22)
= ( D e + D w ) p + D e E + D w W
Therefore, D, the main diffusion coefficient needs to be evaluated at
the faces of each control volume.
The cross diffusion term is written as
jk
1
jk
D c = --- -------- Jg ------- ,j k
J k
j
(3-23)
21
3-9
Source Term
21
Dc
21
21
J d d d = Jg ------ Jg ------
e
w
21
By defining D c
(3-24)
21
= --- Jg and assuming
4
1
ne = --- ( p + E + N + NE ) , etc., we have
4
21
Dc
21
J d d d = D e ( N + NE + S + SE )
21
Dw
(3-25)
( N + NW + S + SW )
Source Term
S = S + S
(3-26)
S = Sp p + S U
(3-27)
Sp = S
3-10
(3-28)
SU = S
(3-29)
Finite Difference
Equation
ap p = a W W + aE E + a S S + a N N + aL L + aH H
+a SW SW +a SE SE +a NW NW +a NE NE
+a LS LS +a LN LN +a HS HS +a HN HN (3-30)
+a WL WL +a WH WH +a EL EL +a EH EH
+S U
The coefficients aP, aW, etc. are known as link coefficients, and the
above equation is known as a finite difference equation (FDE). This
equation is the discrete equivalent of the continuous flow transport
equation we started with.
This FDE, in general, is nonlinear because the link coefficients themselves are functions of p, W, etc. When an FDE is formulated for
each computational cell, it results in a set of coupled nonlinear algebraic equations. No direct matrix inversion method is available to
solve a set of nonlinear algebraic equations. Therefore an iterative
procedure is employed in CFD-ACE at every time step. A linear FDE
is formed by evaluating the link coefficients with the values of
available at the end of the previous iteration.
k
k+1
app
CFD-ACE Theory Manual
k+1
= anb nb
+S U
(3-31)
3-11
Here, the compact notations anb and nb are used to represent the link
coefficients and the values of the flow variable corresponding to the
neighboring grid points. The superscripts k and k+1 denote the previous and current iteration numbers respectively. When the linear set
equation 3-31 is solved, we have an improved estimate for . This
improved estimate is used to update the link coefficients aP, anb and
SU and the linear set is solved again. The iterative procedure is
repeated until a converged solution is obtained. The solution procedures used to solve the linear set are discussed in chapter 8.
Pressure
Gradient Term
The pressure and velocity fields are strongly coupled in fluid flows.
For this and other reasons that will become obvious to the user in the
next chapter, the pressure gradient term appearing in the momentum
equations is treated differently from other source terms.
Let us consider the pressure gradient term of the x-momentum equation. In a BFC system, it is represented as
p
p
p
p
------ = ------ x + ------ x + ------ x
(3-32)
x = i
2
x = i
(3-33)
x = i
equation 3-32 may be represented in a compact form as shown below.
p k
p
------ = -------- ( i )
x
k
3-12
(3-34)
p
------ Jd d d =
x
------
( i) d d d
1x
(3-35)
2x
3x
= A p + A p + A p
where, A1x, A2x and A3x represent projected areas of control volume
faces, e. g., A1x is the area of face of the control volume projected
normal to the x-direction. p, p and p represent pressure differences which, when evaluated at the cell center, are given by
p = pe pw
p = pn ps
(3-36)
p = ph pl
where pe, ps etc. are cell-face pressures that need to be interpolated
from neighboring nodal pressures. Equation 3-35 may be rewritten in
the following compact form
p
kx
------ = A k p
xp
(3-37)
Here, p/x is the discrete equivalent of the continuous pressure gradient p/x. Similar expressions can be obtained for p /y and p /
z.
3-13
3-14
Velocity-pressure Coupling
Introduction
Continuity and
Mass Evaluation
1 J + 1--- ------
k = m
--- --------
J
J t
k
(4-1)
p p
------------------------------------- + G e G w + G n G s + G h G l = m p (4-2)
t
Where Gs represent the mass fluxes through control volume faces as
already mentioned in the previous Chapter. Let us consider the mass
flux at the east cell face
4-1
J
1
G e = ----- U
h1
e
(4-3)
Because the fluid density and velocities are available only at cell centers, Gs need to be interpolated from cell-center values. Linear interpolation of Gs decouples the velocity and pressure fields giving rise
to the well known checkerboard instability, in particular, for incompressible flows. In CFD-ACE, the checkerboard problem is circumvented by following the method proposed by Rhie and Chow (1983)
and later refined by Peric (1985). In this method, the cell-face mass
flux is evaluated by averaging the momentum equation to the cell
faces and relating the cell face velocity directly to the local pressure
gradient. This procedure is presented below.
Referring to figure 3-1 or figure 3-2, the discretized x-momentum
equation for cell P can be represented as
p
a p u p = { a nb u nb + S U } p p ------
x p
(4-4)
where the subscript p after the term in vertical braces indicates that
the link coefficients correspond to the cell P. Dividing equation 4-4
by ap yields
p
u p = a u + S U d p ------
x p
nb nb
(4-5)
a nb
a nb = ------ap
(4-6)
S
S U = -----Uap
(4-7)
where
4-2
p
d p = ------ap
(4-8)
p
u e = a nb u nb + S U d e ------
x e
(4-9)
1
d e = --- ( d p + d E )
2
1
[ { a nb u nb + S U } p
a nb u nb + S U = -2
(4-10)
(4-11)
+ { a nb u nb + S U } E ]
Even though arithmetic averages are used in the two preceding equations for ease of explanation, linear interpolation should be used in the
evaluating of cell-face terms. The terms on the right hand side of
equation 4-11 can be obtained using equation 4-5.
p-
a nb u nb + S U = u p + d p ----x p
(4-12)
4-3
a nb u nb + S U
p-
= u E + d E ----x E
E
(4-13)
1
- ( u + uE )
a nb u nb + S U = -2 p
e
(4-14)
1
+ --- ( d p p
------ + d p
-
2
E ---- x p
x E
The pressure gradient ( p x ) e is evaluated at the cell face following a procedure similar to the one outlined in the previous Chapter.
Note that pressure difference p , which will appear in the evaluation of ( p x ) e , can be directly evaluated from nodal pressures.
{ p } e = p E p p
(4-15)
1
1
p
p
p
u e = --- ( u p + u E ) --- ( d p + d E ) ------ d p ------ d E ------
x e
x p
x E
2
2
(4-16)
It is clear from the preceding equations that the cell face velocity
equals the average of the cell-center velocities plus a correction
involving a third-order pressure gradient term. The pressure terms
serves as a mechanism to avoid the checkerboard problem.
4-4
The other two velocity components at the east cell face, e and we,
are similarly evaluated and the contravariant velocity component at
the east face is evaluated from
1
U e = u e F 1x + e F 1y + w e F 1z
(4-17)
with
F 1x = h 1 i
1
F 1y = h 1 j
1
F 1z = h 1 k (4-18)
1
P if U 1e 0
UP
Upwind: e =
1
E if U e < 0
CD
Central: e
1
= --- ( P + E )
2
(4-19)
(4-20)
e = e
UP
+ ( 1 )e
(4-21)
4-5
Pressure
Correction
Equation
SIMPLEC
Algorithm
kx
a p u p = a nb u nb + S u A p k P p
(4-22)
with the subscript P again indicating that the equation is written for
cell center P. The pressure field should be provided to solve equation
4-22 for u. However, the pressure field is not known a priori. If the
preceding equation is solved with a gussed pressure (or the latest
available pressure in an iterative procedure) p*, it will yield velocity
u* which satisfies the following equation.
a p u p = { a nb u nb + S u } p A p k P p
kx
(4-23)
In general, u* will not satisfy continuity. The strategy is to find corrections to u* and p* so that an improved solution can be obtained.
Let u and p stand for the corrections.
4-6
u = u + u
(4-24)
P = P + P
(4-25)
a p u p = { a nb u nb } A p k P p
p
(4-26)
u p = d kp k P p
(4-27)
where
kx
k
dp
Ap
= ----------------------a p a nb
(4-28)
(4-29)
p p
------------------------------------- + G e G w + G n G s + G h G l = m p (4-30)
t
4-7
If u* and p* are used to evaluate Ge, Gw, etc., the resulting G*e, G*w,
etc. will not satisfy the preceding equation. Let G represent the correction to G, i.e.,
G = G + G
(4-31)
p p
--------------- + Ge Gw + Gn Gs + Gh Gl = Sm
t
(4-32)
(4-33)
( G e G w + G n G s + G h G l )
By using equation 4-31 we get an expression for G e
J
1
J
1
Ge = ----- U + ----- U
h
(4-34)
The user should note that in deriving the preceding equation, terms
containing products of primes have been neglected.
For incompressible flows is zero. For compressible flows is
estimated as
= ------ p
p
4-8
(4-35)
1
------ = ------p
RT
(4-36)
Ue
= ue F 1x + v e F 1y + we F 1z
(4-37)
where F1x, etc. are obtained from equation 4-18; equation 4-29 is used
to obtain u e and v e and w e are obtained similarly starting with yand z-momentum equations following the procedure outlined in this
section.
Expressions for the contravariant velocity corrections and fluid density corrections at all the cell faces are obtained and substituted in
equation 4-32 yielding an FDE for the pressure correction
ap Pp = aW PW + aE PE + aS PS + aN PN
+a L P L + a H P H + S m
(4-38)
1.
2.
3.
4.
4-9
5.
6.
7.
8.
4-10
Special Topics
Introduction
Moving Grid
Implementation
5-1
d
----dt
d = Vg ds
(5-1)
The continuity and the generalized transport equations for the problem are written in integral form as
d
----dt
d
----dt
+ ( ( V V g ) ds ) = 0
(5-2)
+ ( V V g ) ds = q ds +
s
S d
(5-3)
o
o
-------------------- = V g ds + ( 1 ) V g ds
t
face
f
(5-4)
where stands for volume of the control volume at a new time step
and superscript o stands for values at the old time step. Note that q =
1.0 is the standard Euler differencing and q = 0.5 is the Crank-Nicolson scheme and the value of q should be the same as the one used for
5-2
other conservation equations for consistency. For numerical purposes, it is convenient to define a volumetric grid flux at a control volume (or cell) face, f, as
f =
Vg ds
(5-5)
o
-------------------- = f + ( 1 ) f
t
face
(5-6)
If stands for the volume swept by the cell face f, during one time
step, it yields from the basic geometric requirement that
o
-------------------- = f
t
f
(5-7)
Hence, a natural way for calculating the volumetric grid flux is to let
f
---------- = f + ( 1 ) o f
t
(5-8)
o
1
f = -----------f + 1 --- f
(5-9)
or
G f = f f
(5-10)
which when subtracted from the cell face mass flux, Gf, results in the
effective mass flux that should be used in the evaluation of convective
CFD-ACE Theory Manual
5-3
Conjugate Heat
Transfer Analysis
5-4
Ice
Glass
PVB
Glass
Main
Computation
Domain
Vair
Air Entrained
Defroster
Inlet
Interface
Constraints
Heat flux at the interface must be equal both on solid and fluid
sides (see figure 5-2).
Temperature has to be continuous across the interface.
5-5
f kf Cps
2
Solid
1
T2i
Hs
q
Tfi
f kf Cpf
Hf
Tsi
T1i
Fluid
K s ( T ) s n = K f ( T )f n
(5-11)
T i = T si = T fi
(5-12)
where the subscripts s and f stand for solid and fluid respectively. The
subscript i indicates interface. The user should note that neither
energy nor enthalpy are continuous across the interface due to the difference in specific heat Cp.
Equivalent Thermal
Conductivity
5-6
K s T s n = K f T f n = K e T n
(5-13)
T n is expressed as
2
3
T e 1
T e
T e
T n = ------ --------------- + ------ --------------- + ------ --------------- n
e 1 e e 2 e e 3 e
1
(5-14)
where e1, e2, and e3 are the covariant base vectors in , , and directions, while e1, e2, and e3are the contravariant base vectors.
By assuming
T s T s T i T s T i1 T i2 T s T i3 T i4
-------- = ---------------- , -------- = --------------------- , -------- = --------------------- (5-15)
s
s
s
in the solid phase, and
T
T i T f T f T i1 T i2
, -------- = --------------------- ,
--------f = --------------
f
f
T
T i3 T i4
-------s- = -------------------
f
(5-16)
AT f + BT s + ( D C ) + ( F E )
T i = -------------------------------------------------------------------------A+B
(5-17)
where
5-7
Ks
1
B = --------- ------------
s e 1 n s
Kf
1
A = -------- ------------
f e 1 nf
K f ( T i1 T i2 ) e 2 n
K s ( T i1 T i2 ) e 2 n
C = -------------------------------- ---------------- D = -------------------------------- ---------------
e2 e
e2 e
2 f
2 s
K f ( T i3 T i4 ) e 3 n
K s ( T i3 T i4 ) e 3 n
E = -------------------------------- --------------- F = -------------------------------- ---------------
e3 e
e3 e
3 f
3 s
Note that Ti1, Ti2, Ti3 and Ti4 are the interfacial temperatures at the
corners. By substituting equation 5-14 into equation 5-13, the equivalent thermal conductivity Ke is obtained after some tedious manipulation.
AB ( T s T f + A ( D + F ) + B ( C + E ) )
K e = ---------------------------------------------------------------------------------------( A + B )[ G( T s Tf + H + 1 ) ]
(5-18)
where
1
1
G = ------------------------ ------------
s + f e 1 n
T i1 T i2 e 2 n
H = --------------------- ---------------
e 2 e
2
T i3 T i4 e 3 n
I = --------------------- ---------------
e 3 e
3
Turbulent Flow
Considerations
5-8
f Cpf u ( Tf Ti )
qw = Ks T nw = --------------------------------------+
T
(5-19)
w 2
u = -------
(5-20)
(5-21)
f C pf u
K f ---------------------+
T
(5-22)
or
where is the normal distance between the wall and the first cell.
5-9
5-10
Boundary Conditions
Introduction
The incorporation of boundary conditions into the CFD-ACE finitevolume equations is discussed in this chapter. It is important to take
note of the boundary values that are required by the program and the
way in which those values may be derived from the boundary conditions specified by the user. First, the general treatment of boundary
conditions will be presented and then the treatment of each type of
boundary condition used in the program will be discussed.
General
Treatment
a P P = a + a + a + S
E E
N N
S S
(6-1)
with coefficient a W having been set to zero. The links to the boundary node are incorporated into the source term S, expressed in linearized form as
S = SU + S P
P
(6-2)
6-1
Inlet Boundaries
P
B
S
Figure 6-1. Boundary Cell
Inlet Boundaries
CFD-ACE provides for two types of inlet boundary conditions; a prescribed mass flux boundary and a prescribed total pressure boundary.
These are discussed below.
Prescribed Mass
Flux Inlet
The mass flux boundary condition is applied to the pressure correction equation by setting the mass flux at the boundary as
GB = B uB A
(6-3)
6-2
Inlet Boundaries
A
SU = SU + B uB A + ------- B
A
Sp = Sp B uB A + -------
(6-4)
(6-5)
Prescribed Total
Pressure Inlet
At this kind of inflow boundary, the total pressure ( P o ) and total temperature ( T o ) are specified. CFD-ACE offers two different
approaches to obtaining boundary values of velocity components and
static pressure:
Velocity Extrapolation
6-3
Inlet Boundaries
V B
P o = P B + -----------2
(6-6)
where V B is the magnitude of the extrapolated velocity. For compressible flows, assuming that a perfect gas undergoing an isentropic
process we have,
1 2
P o = P B 1 + ----------- M
2
----------1
(6-7)
1 2
T o = T B 1 + ----------- M
2
(6-8)
M = V B RT B
(6-9)
Pressure Extrapolation
V = VBn
6-4
(6-10)
Exit Boundaries
here n is the unit vector that specifies the inflow direction. The
inflow direction needs to be specified by the user. This can be done
by setting a vector input in the global Cartesian coordinate system, in
a cylindrical x r system or simply assuming that the inflow
direction is normal to the local boundary cell face.
With the boundary velocities and static pressure thus determined, the
source terms for the equations are then constructed in the same manner as for the prescribed mass flux inlet except that the boundary
static pressure, PB, is used in the momentum equations.
Exit Boundaries
Prescribed Pressure
Exit
At a prescribed pressure exit, the pressure correction equation is modified to include a link to a boundary pressure correction value of zero.
This can be accomplished simply by modifying SP
Sp = Sp aB
(6-11)
A
SU = SU + ------- P
A
Sp = Sp ------
(6-12)
(6-13)
6-5
Exit Boundaries
When the mass flow is into the domain the source term modification
are
A
SU = SU + m + ------- B
(6-14)
A
S p = S p m + -------
(6-15)
if the mass flow is into the domain. Note that when flow is into the
domain then the boundary value B is convected into the domain,
This makes it important that reasonable boundary values be provided
for all dependent variables.
Extrapolated
Conditions Exit
SU = SU p uA
(6-16)
Sp = Sp p uA
Combined Fixed
Pressure and
Extrapolated Exit
6-6
(6-17)
Symmetry Boundaries
Symmetry
Boundaries
Wall Boundaries
Momentum
Equations
SU = SU + DB
(6-18)
Sp = Sp D
(6-19)
------D = A
(6-20)
or as
A ( up uB )
D = ---------------------------- ( up uB )
w A
= --------------------( up uB )
(6-21)
6-7
Wall Boundaries
Enthalpy Equation
The source term in the enthalpy equation is modified in the same fashion as for the momentum equations. In this case the diffusive coefficient D can be expressed as
qw A
D = ---------------------( Tp TB )
(6-22)
Where, qw is the conductive heat transfer to the wall and it is evaluated from finite differences for laminar flows and by wall functions
for turbulent flows. Determination of wall temperature depends upon
the kind of thermal boundary condition applied. CFD-ACE allows for
almost all kinds of generally applied thermal boundaries, including
constant temperature, constant heat flux, adiabatic wall, external convection and radiation. The methods in which T B are calculated are
outlined as follows:
stant)
Constant heat flux: in this case, the net heat flux across the
boundary is a specified constant, namely,
K eff ( T B T P ) ( R 1 R 2 T B ) = q
(6-23)
R 1 + q + K eff T P
T B = ------------------------------------K eff + R 2
6-8
(6-24)
Wall Boundaries
K eff ( T B T P ) ( R 1 R 2 T B ) = h c ( T ext T B )
4
4
+ ( T T B )
where the additional symbols, h c, T ext, and T are userdefined external heat-transfer coefficient, ambient temperature,
surface emissivity and temperature of radiation source/sink on
the exterior of the computational domain, respectively. is the
Stefan-Boltzmann constant. From equation (6-25) we have,
4
R 1 + q + K eff T P + h c T ext + T
T B = ----------------------------------------------------------------------------------3
K eff + R 2 + h c + T B
here T B is the wall temperature at the previous iteration.
Turbulence
Equations
SU = 10
20
Sp = 10
20
(6-25)
6-9
6-10
Wall Boundaries
Introduction
Solution Methods
In chapter 4 it was pointed out that CFD-ACE uses an iterative, segregated solution method wherein the equation sets for each variable are
solved sequentially and repeatedly until a converged solution is
obtained. In this chapter, the overall solution method is presented in
flowchart form or both the SIMPEC and PISO solution algorithms.
Additional details of the solution procedure are also presented,
namely under-relaxation, use of block correction acceleration and relevant details of the linear equation solvers.
Overall Solution
Procedures
SIMPLEC
Algorithm
7-1
The steps involved in assembling and solving the equation set for a
particular dependent variable are shown in figure 7-2. The calculation of the source terms and link coefficients as well as the application
of the boundary conditions have already been discussed in chapters 4
and 7. The other operations shown in figure 7-2 will be discussed in
sections, Under-Relaxation, on page 8-5 through Linear Equation
Solvers, on page 8-7.
7-2
At t = tf
Prescribe Initial Flow
t = t + t
Solve u, v, w
Repeat
NSTEP times
Solve p
Correct p, u, v, w
Repeat
NITER times
Solve h
STOP
7-3
Apply Under-relaxation
PISO Algorithm
7-4
At t = tf
Prescribe Initial Flow Field
t = t + t
Solve k,
Repeat NITER
times
Repeat NSTEP
times
Solve Compositions
YES
Steady?
NO
STOP
7-5
Under-relaxation
Under-relaxation
Under-relaxation of dependent and auxiliary variables is used to constrain the change in the variable from iteration to iteration in order to
prevent divergence of the solution procedure. For all dependent variable except the pressure correction, this is done by modifying equation 3-31 in the following way
ap ( 1 + I )p = anb nb + SU + ap Ip
(7-1)
where
p is the current value of p . At convergence, when there is
no change in p from one iteration to the next, the equation is not
modified at all by the addition of these terms. Prior to convergence,
however, they provide a link between the new value of p and the
current iteration value
p . The larger this link or the value of I the
stronger the under-relaxation.
Reasonable values for I can be deduced if we think of it as being analogous to the transient term link coefficient
I = -----------tf ap
(7-2)
where tf is a false time step. If I is equal to 1.0, then tf is comparable to the maximum time step allowed in an explicit time-marching solution scheme. Since CFD-ACE uses an implicit solution
scheme it is possible to use larger false time steps and, hence, smaller
values of I. Values of I in the range of 0.2 to 0.8 are common. In very
difficult problems it may be necessary to use values of I in excess of
1.0.
The auxiliary variables , P, T and v can be under-relaxed by specifying a linear under-relaxation factor, , which is applied in the following way
7-6
Block Correction
new
= + ( 1 )
(7-3)
Block Correction
a p ijk = a W i 1, j, k + a E i + 1, j, k + a S i, j 1, k
ijk
i jk
ij k
ijk
+a N i, j 1, k + a L i, j , k 1
ijk
ij k
(7-4)
+a H I, J, K + 1 S U
ijk
i jk
ijk = * ijk + i
(7-5)
7-7
a p i = a W i 1 + a E i + 1 + SU
i
i
i
i+1
(7-6)
a p = ( a p a S a N a L a H )
i
i jk
ijk
ij k
i jk
ijk
k j
(7-7)
aW = a W
i
ijk
k j
(7-8)
aE = aE
i
ijk
k j
(7-9)
where
S U = ( a W i 1, j, k + a E i + 1, j, k
i
ijk
i jk
k j
+a S i, j 1, k + a N i, j 1, k
ij k
ijk
+a L i, j, k 1 + a H i, j, k + 1
ij k
ij k
+SU
i jk
a p
i jk
ijk
(7-10)
Linear Equation
Solvers
7-8
Once the equation set for a variable has been assembled it must be
solved. Iterative equation solvers are preferred for this task because
[ A ] = S
(7-11)
aP
aE
aN
aH
aW
A=
aS
aL
7-9
[ A + B ] = S + [ B ]
(7-12)
The matrix B must be such that A + B is easily factored into upper and
lower triangular matrices L and U. L and U should have no more nonzero entries than the upper and lower halves of the original coefficient
matrix A making them as easy to store as A. The modified equation
then becomes
[ L ] [ U ] = S + [ B ]
(7-13)
[ L ]V = S + [ B ]
(7-14)
V = [ U ]
(7-15)
where
k+1
from [ L ]V
k+1
from [ U ]
Evaluate
k+1
k+1
= S + [ B ]
= V
k+1
Conjugate Gradient
Squared Solver
7-10
into the CFD-ACE program. The CGS algorithm applied to the system A = S is expressed as follows:
Initialization (n = 0)
r o = S A o
q o = p 1 = 0 ; p 1 = 1
Iteration (n > 0)
n
T
n = r o r n ; n = -----------n 1
u n = r n + n qn
pn = un + n ( qn + np
)
n1
v n = Ap n
T
n = r o v n ; n = -----nn
qn + 1 = u n n v n
rn + 1 = r n n A ( u n + qn + 1 )
n + 1 = n + n ( un + q n + 1 )
The convergence rate of conjugate gradient algorithms depends on the
spectral radius of the coefficient matrix and can be effectively accelerated by preconditioning the system. This preconditioning is accom-
= p 1 A
plished by transforming the system A
to
7-11
7-12
Turbulence
Introduction
Overview
t ql .
(8-1)
8-1
Chapter 8: Turbulence
Overview
flows. Two equation models solve transport equations for two quantities used to estimate q and l.
The two-equation models in CFD-ACE can be further divided into
high Reynolds number and low Reynolds number models. Here the
qualifier high Reynolds number refers to the turbulent Reynolds
number,
k2
Re t = ----- ,
v
(8-2)
8-2
yu
= -----------
(8-3)
Baldwin-Lomax Model
Chapter 8: Turbulence
u
= ----u
(8-4)
w 1 2
,
u = ------
(8-5)
= y .
(8-6)
= a ln y + b .
(8-7)
This region is, therefore, also known as the logarithmic layer. The
layer between y+ = 5 and y+ = 30 is commonly known as the buffer
layer.
The following sections describe the turbulence models available in
CFD-ACE. In the description of models, the overbar for and , and
tilde for u, v, etc. have been dropped for convenience.
Baldwin-Lomax
Model
8-3
Chapter 8: Turbulence
Baldwin-Lomax Model
( t ) inner for y y c
t =
( t ) outer for y y c
(8-8)
where y is the normal distance from the wall and yc is the crossover
point between the inner and outer layers where the values for the eddy
viscosity using the expressions below are equal.
The eddy viscosity in the inner layer is obtained as
(t)
inner
= l ,
(8-9)
l = y [ 1 exp ( y A ) ] ,
(8-10)
and the velocity scale is modeled as the product of l and the vorticity
magnitude, = u ,
q = l ,
(8-11)
8-4
Chapter 8: Turbulence
( t )outer = KF w F k(y) ,
(8-12)
where K is a constant,
2
(8-13)
and
1
F k ( y ) = -----------------------------------------6
Ck y
---------1 + 5.5
y max
(8-14)
U dif =
The values used for the constants appearing in the outer layer eddy
viscosity are
C k = 0.3,
C w = 0.1,
K = 0.0269.
Standard k-
Model
The k- model is a two equation model that employs partial differential equations to govern the transport of the turbulent kinetic energy, k,
and its dissipation rate, . Several versions of the k- model are in use
8-5
Chapter 8: Turbulence
C 3 / 4 k 3 / 2
l = --------------------- .
(8-16)
C k2
v t = ----------- .
(8-17)
(8-18)
+ ------xj
(8-19)
+ ------ ------ xj
u u 2 u m u 2 u m
P = vt -------i + -------j --- ---------- ij -------i --- k --------- xj xi 3 x m x j 3 x m
8-6
(8-20)
Chapter 8: Turbulence
C = 0.09,
C = 1.44,
1
C = 1.92,
2
k = 1.0,
= 1.3.
The standard k- model is a high Reynolds model and is not intended
to be used in the near-wall regions where viscous effects dominate the
effects of turbulence. Wall functions are used in cells adjacent to
walls. The k and transport equations are not numerically integrated
in these cells. Instead, semi-empirical expressions are used to relate
k, , and the friction velocity u . These expressions are obtained
from analysis of the momentum and turbulence equations for a flat
plate boundary layer, assuming a logarithmic velocity profile.
2
u
k = ---------C
3--4
C
(8-21)
3--2
k
= --------------y
(8-22)
= y
for y
< 11.5 ,
(8-23)
8-7
Chapter 8: Turbulence
1
+
+
= --- ln ( Ey ) for y > 11.5 .
(8-24)
The wall shear stress is calculated iteratively from the known values
of y and u in the first cell. The constants appearing in equation 8-24
are experimentally determined to be E=9.0 and = 0.4. Because the
semi-empirical relations for k and in the first cell assume a logarithmic velocity profile, the turbulence wall functions are strictly valid
only if the center of the cell nearest the wall is inside the logarithmic
+
2
1
2
Cp ( T Tw ) + --- eff ( V Vw )
2
= -------------------------------------------------------------------------- ,
q w u
(8-25)
= u
+
= t ( u + P )
for
y yT ,
for
y > yT ,
(8-26)
(8-27)
RNG k- Model
Chapter 8: Turbulence
1
---
t 4
+
P = 9 ----- 1 ----- ,
(8-28)
and y T is the value of y+ at which both the expressions in equation 826 and equation 8-27 yield the same value of T+. If = 0.7 and t =
+
0.9, then P+ = 2.13 and y T = 9.585. The turbulent heat flux at the
wall is evaluated with equation 8-25, using data from the cell nearest
the wall.
RNG k- Model
C = 0.085
C = 1.68
2
k = = 0.7179
The coefficient C becomes a function of , the ratio of time scales
1
for turbulence and mean strain rate.
1 ------
0
C = 1.42 -----------------------3
1
1 +
(8-29)
8-9
Chapter 8: Turbulence
Sk
= ------ , S =
2S ij S ji
(8-30)
The constants in equation 8-29 have the values 0=4.38 and =0.015.
The rate of mean-strain tensor, S ij , is defined as follows,
1 u u 2
S ij = --- -------i + -------j --- V ij
2 x j x i 3
(8-31)
Kato-Launder
k- Model
U i
P = 2V t S = 2V t S ij
x
(8-32)
U
P = 2sV t S = 2V t 2S ij
xj
(8-33)
8-10
k- Model
Chapter 8: Turbulence
1 U U j 2
ij = --- i
--- V ij
2 xj xi 3
k- Model
(8-34)
The k- model is a two equation model that solves for the transport of
, the specific dissipation rate of the turbulent kinetic energy, instead
of . The k- model in CFD-ACE is based on Wilcox (1991). The
eddy viscosity in this model is
k
t = C ---- .
(8-35)
( ) +
( u ) = C ----------- C
1
2
k
t
xj
+
(8-36)
(8-37)
t
+ -------
xj
x j
The model parameters in the above equations are all assigned constant
values.
k = 0 at y = 0 ,
(8-38)
8-11
Chapter 8: Turbulence
= 7.2 ----- at y = y 1 .
2
y
(8-39)
where y 1 is the normal distance to the wall of the center of the cell
next to the wall. The location of the cell center should be well within
+
Low Reynolds
Number k-
Model
----------( uj k ) =
( k ) +
xj
t
xj
t k
+
------ -------
k xj
(8-40)
+ (P D)
t
P
x
1
xj
k
t
xj
j
(8-41)
2
C f 2 -------- + E,
2
k
8-12
Chapter 8: Turbulence
k
t = C f --------
(8-42)
(8-43)
2
+
D = 2vk y ; E = 2v ----- exp ( 0.5y )
2
y
The wall shear stress is computed from finite differences for this
model. Therefore, the first grid-point should be placed in the laminar
+
8-13
Chapter 8: Turbulence
8-14
Reaction Models
Basic Definitions
And Relations
Calculation of reactive flow requires the consideration of both stoichiometry and reaction kinetics. Stoichiometry is the description of the
conservation of mass and elements. Reaction kinetics is the description of the individual steps that make up a chemically reacting system
and the specification of the rates at which those steps progress.
A distinction will be made between elementary and global reactions.
A global reaction is one such as
CH4 + 2 O2 CO2 + 2 H2O
which is correct in the stoichiometric sense, because all elements are
conserved. This global step does not describe the true path of meth-
9-1
Composition
variables
Several different composition variables are used for flow with mixing
or reaction. The mass fraction of species i in a multicomponent system, Yi, is defined as the mass of the ith species per unit mass of the
mixture. Similarly, the mole fraction xi is defined as the number of
moles of the ith species per mole of the mixture. The mole and mass
fractions are related to each other by the molecular weight of the ith
species, Mi, and the mixture molecular weight, M.
Yi M
x i = --------Mi
(9-1)
1
M = -------------------------- =
( Yi Mi )
xi Mi
(9-2)
Y
c i = --------i
Mi
(9-3)
9-2
Y
n i = -----iMi
(9-4)
pi = xi p
Chemical rate
expressions
(9-5)
N sp
ij i =
i=1
ij i , j = 1, , N rxn ,
(9-6)
i=1
where i is the chemical symbol for species i, ij and ij are the forward and reverse stoichiometric coefficients for the ith species in the
jth reaction. Equation 9-6 can be written more compactly as
N sp
ij i = 0 , j = 1, , N rxn ,
(9-7)
i=1
9-3
i1
i2
i1
i2
H2
H2O
CO
CO2
1
0
0
0
0
1
0
0
O2
N2
i =
ij q j
(9-8)
j=1
The rate-of-progress variable for the jth reaction, qj, can be generally
expressed as
N sp
qj = ( k f )
j
i=1
N sp
ij
ci
( kr )
ij
ci ,
(9-9)
i=1
reactions, ij = ij and ij = ij , where ij and ij are the stoichiometric coefficients defined in equation 9-6. Reactions with this
behavior are said to follow the law of mass action. The concentration
exponents in equation 9-9 are not necessarily related to the stoichiometric coefficients for global reactions.
The rate coefficients are assumed to have an Arrhenius form
n E a RT
k f = AT e
(9-10)
k
k r = -----fKc
(9-11)
sp
sp p
gi
0
ln ( K c ) = ij ln ------- ij -------
j
RT
RT
i = 1
i=1
(9-12)
9-5
Elementary reactions are sometimes written with a third body, usually designated with the symbol M, which can be any species. For
example,
H + O 2 + M = HO2 + M
The rate-of-progress variable for these reactions is
N
sp
sp
sp
i j
i j
ij
q j = ij c i
( kf ) ci ( kr) ci
j
j
i = 1
i=1
i=1
Mixture fractions
(9-13)
fk
------ + ( m k )
evap
xj
(9-14)
In the preceding equation fk represents the mixture fraction for the kth
mixture. Note that this equation contains no source terms due to
chemical reaction. The only source term is due to the evaporation of
spray droplets. The diffusion coefficient is the same for all mixture
fractions.
Mixture fractions are normally associated with one or more inlet
boundaries and normalized such that the value is 1 for the boundaries
associated with that mixture and 0 for other boundaries. A mixture
9-6
Y
---- ( Y i ) + ------- ( u j Y i ) = ------- --------i + ( m i )
evap
t
xj
xj xj
(9-15)
where
Yi =
ik fk
(9-16)
9-7
Mixture fractions are also used with certain reaction models to calculate the composition of reacting flows. These models are discussed
below.
Instantaneous
Chemistry Model
( Yi ) =
ik f k
(9-17)
k=1
1 1
2 2
i i
(9-18)
i=1
9-8
i M i
r i = -----------1M1
(9-19)
Note that the stoichiometric coefficients in equation 9-19 are for the
overall reaction and, therefore, positive for product species and negative for fuel and oxidizer. Positive values of ri indicate production
and negative values indicate consumption. The instantaneous reaction consumes either all the fuel or all the oxidizer, whichever is limiting. The amount of fuel that is consumed is
u
( Y 1) ( Y2 )
Y 1 = min -------------, -------------
r1 r2
(9-20)
Y i = ( Y i ) + r i Y 1
(9-21)
The right-hand side of equation 9-21 is only a function of the K mixture fractions. K 1 transport equations must be solved for the mixture fractions. These equations have no source terms due to chemical
reactions.
9-9
Single-Step Finite
Rate Chemistry
Model
In the finite-rate chemistry model, as the name implies, a single reaction proceeds at a finite rate. The reaction stoichiometry is specified
in the same manner as in the instantaneous chemistry model (equation
9-18). The model is restricted to two reactant species. In addition to
the stoichiometry, a rate expression must be specified. The primary
difference between the finite-rate and instantaneous models is that the
mass fraction of fuel is calculated by solution of a transport equation
with a source term due to chemical reaction for the finite-rate model.
The mass fractions of the other species are calculated from the mixture fractions and the mass fraction of fuel.
The molar production rate of species i due to the single-step reaction
is
1 2
i = i kf c1 c2
(9-22)
An Arrhenius form (equation 9-10) is used for the reaction rate coefficient. The reaction is irreversible (i. e., the reverse rate coefficient is
zero). As this is a global model, the concentration exponents do not
have to be the same as the stoichiometric coefficients.
The transport equation for the mass fraction of fuel, Y 1 , is
Y 1
(9-23)
evap
Y i = ( Y i ) + r i Y 1
9-10
(9-24)
Equilibrium
Chemistry Model
G =
i=1
nii =
i=1
n i i + RT ln ----i-
p 0
(9-25)
9-11
a ji n i = b j for j = 1, , M
(9-26)
i=1
= G + j a ji n i b j
j = 1 i = 1
M
(9-27)
= i
a ji j for i = 1, , N
ni
j=1
(9-28)
= bi
a ji n j for i = 1, , M
i
j=1
(9-29)
9-12
p
ln c i = -------i + ln ------o- +
RT
RT
j
a ji ------- for i = 1, , N (9-30)
RT
j=1
and
N
a ji c i = b j for j = 1, , M
(9-31)
i=1
defining
po
j
j
R j = ------- exp ------ and Z j = exp ------- .
RT
RT
RT
Equation 9-30 is simplified as
M
ci = Ri
( Zj )
a ji
for i = 1, , N .
(9-32)
j=1
ak i
a R
(
Z
)
ji i k b j = 0 for j = 1, , M . (9-33)
i = 1
k=1
N
Thus we have M nonlinear algebraic equations for the unknown values of Zj. The values of bj are calculated from the mixture fractions
using equation 9-16, giving
N
bj =
i = 1k = 1
a ji ik f k
------------------
Mi
(9-34)
9-13
The multistep reaction model does not use the concept of mixture
fractions that are used in the other chemistry models. Transport equations are solved for the mass fraction of Nsp species. The transport
equation for species i is
(9-35)
(Si )
n+1
= Mii + Mi
j=1
i
n+1
n
Y )
(Y
Y
j
(9-36)
where the indices n and n + 1 denote the iteration at which the corresponding quantity is evaluated. There are two methods available for
the solution of equation 9-35. The first uses the full Jacobian array in
equation 9-36 and couples the solution of all mass fractions in a pointiterative equation solver. The other method only uses the diagonal
elements of the Jacobian array and solves each mass fraction equation
sequentially with a whole field equation solver.
9-14
This chemistry model cannot be used with liquid spray because the
mass source terms due to evaporation are not included in the transport
equations.
Surface Reaction
Models
The surface reaction models allow the calculation of deposition, etching, or catalytic reaction at surfaces. The surface reaction provides a
boundary condition for the mass fractions of species in the fluid,
rather than a source term in the transport equations. The mass flux of
reacting species to the surface (or away from the surface for species
produced by the reaction) equals the rate at which the species is consumed (or produced) by the reaction on the surface.
J in = ( i )
S
(9-37)
RT W
( i ) = k ( T W )Y AW -------------S
2M A
(9-38)
( i )S = k ( T W ) [ A ] [ B ]
(9-39)
9-15
Turbulencecombustion
Interaction
Turbulence-combustion Interaction
( Y 1, , Y N ;x, t ) = ( Y 1, , Y N )
( P ( Y 1, , Y N ;x, t ) dY 1 dY N )
(9-40)
( Y 1, , Y N )P ( Y 1, , Y N )
P (Y 1, , Y N) = ----------------------------------------------------------------- ( Y 1, , Y N )
(9-41)
9-16
Turbulence-combustion Interaction
Determination of
PDF
( Y f, max Y f )
= ----------------------------------------( Y f, max Y f, min )
(9-42)
where Y f is the mass fraction of the fuel in the one step reaction and
the minimum and maximum values are functions of the mixture fraction. The mixture fraction and reaction progress are assumed to be
independent, so the two-dimensional PDF is a product of the two onedimensional PDFs.
P ( f, ) = P ( f )P ( )
(9-43)
The one-dimensional PDFs both have two parameters that are related
to the average and variance of the mixture fraction or reaction
progress. Transport equations are solved for the average and variance
of the corresponding variable. (Note: A transport equation is solved
for the average fuel fraction instead of the average reaction progress
because the reaction progress is not well defined when no fuel or no
oxidizer is present. The average reaction progress is calculated from
the average fuel fraction and mixture fraction.)
The transport equations for the average mixture fraction and average
fuel fraction are derived by averaging equation 9-14 and equation 9CFD-ACE Theory Manual
9-17
Turbulence-combustion Interaction
2
2
f 2
---- f + ------- u j f = ------- --------f- + 2 -------
x
xj xj
t
xj
j
(9-44)
2
CD ------ f
k
2
2
2
-
- 2
------------------------- +
+ 2
u =
xj xj
t xj j
xj
(9-45)
2
CD ------ + S
k
In preceding equation, CD has the value of 2.
Reaction Progress PDF
The PDF for the reaction progress is a discrete function with 3 possible values.
2
P ( = 0 ) = -----
2
P ( = ) = 1 ------------------ ( 1 )
(9-46)
P ( = 1 ) = ---------------( 1 )
Mixture Fraction PDF
9-18
Two choices are available for the mixture fraction PDF: a to-phat
and beta PDF. The to-phat PDF has uniform probability between a
Turbulence-combustion Interaction
Po , f = 0
P ( f ) = c, 0 < a f b < 1
P ,f = 1
1
(9-47)
The parameters for the tophat PDF are given below, as functions of
the average and variance of the mixture fraction.
2
2
2
3 f < min ( f , ( 1 f ) )
P0 = P1 = 0
2
a = f 3 f
2
b = f + 3 f
1
c = ----------------2
2 3 f
2
2
f < 1
--- and 3 f < 2 f 3 f
2
3 f f
P 0 = -----------------------2 2
3 ( f + f )
P1 = 0
a=0
2
3 ( f + f )
b = -----------------------2f
3
8 f
c = ---------------------------2 2
2
9 ( f + f )
9-19
Turbulence-combustion Interaction
f > 1
--- and
2
P0 = 0
2
2
3 f ( 1 f ) ( 3f 1 )
3 f ( 1 f )
P 1 = -------------------------------------2
2
3(f + (1 f ) )
2
( 1 f ) ( 3 f 1 ) 3
a = --------------------------------------------------f2 ( 1 f )
b=1
f
otherwise
3
8 f
c = -----------------------------------------2 2
2
9 ( f + ( 1 f ) )
P 0 = ( 1 f )
2
3 f 1 f + 3 f
2
P 1 = f 3 f ( 1 f ) + 3 f
a=0
b=1
2
c = 6 ( f ( 1 f ) f )
(a + b ) (a 1)
(b 1)
P ( f ) = ------------------------ f
(1 f)
( a ) ( b )
(9-48)
9-20
Turbulence-combustion Interaction
f ( 1 f )
a = f ------------------ 1
2
f ( 1 f )
b = ( 1 f ) ------------------ 1
2
Determination of
Averaged Variables
(9-49)
T =
T ( f, )P ( f, ) df d
(9-50)
0
1P ( f, )
1
--- = ----------------df d
( f, )
(9-51)
Since the mixture fraction and reaction progress are independent variables and the PDF for the reaction progress only has discrete values
the two-dimensional integrals can be evaluated as a sum of onedimensional integrals. For example
T = c 1 T ( f, 0 )P ( f ) df + c 2 T ( f, )P ( f ) df
c 3 T ( f, 1 )P ( f ) df
(9-52)
9-21
Turbulence-combustion Interaction
the transport equations are solved and stored. During the solution of
the transport equations governing the problem, the average values of
variables are determined by linear interpolation from the stored data.
9-22
10
Introduction
Spray Model
Size Distributions
Droplet dynamics
Droplet heat and mass transfer
Droplet source/sink terms due to interactions with the ambient
Droplet turbulent dispersion
Droplet breakup
Droplet collision and coalescence
Size Distributions
10-2
Size Distributions
In the first option, the user specifies a discrete set of size groups as
well as the fractions of injected mass contained in the groups. The
sum of the mass fractions for an injector must add up to unity. For
log-normal and Rosin-Rammler distributions, the mass fractions for
each size group are calculated from the distribution functions. The
user need to specify the volume mean diameter of the droplets, the
standard deviation and the minimum and maximum size of the droplets. From the specified range of the droplet size, a (user-specified)
number of sizes are selected at equal intervals. The mass fraction for
each size group is calculated from the following equations.
Log-normal Distribution
2
D
F = ---- e
D
( ln D ln D )
--------------------------------2
where D is the volume mean diameter of the droplet cloud and is the
standard deviation.
Rosin-Rammler Distribution
D q
----
D
q1
D
F = q -------------- e
q
D
where q is a constant that lies between 1.5 and 4. For both of these
distributions, the fraction of mass contained in size group D is calculated as,
3
FD
f m = ----------------- m
3
FD
where the summation is over all the size groups of an injector and m
is the total mass flow rate through the injector.
10-3
Droplet Dynamics
Droplet Dynamics
d
md ------ = CD ( U ) U
dt
A
-----d- Vd p + md g
2
.,
(10-1)
U d
Re = ----------------------
(10-2)
where is the dynamic viscosity of the gas. The following correlations are used to calculate the drag coefficient:
Incompressible Flow:
24
C D = -----Re
24
0.687
]
C D = ------ [ 1 + 0.15Re
Re
C D = 0.44
10-4
for Re < 1
for 1 < 1Re < 10
for Re < 10
(10-3)
Droplet Dynamics
Compressible Flow:
For compressible flows, two correlations: one for subsonic and one
for supersonic flows, are used to calculate the drag coefficient.
Depending on the local relative Mach number (M), the appropriate
correlation is used.
T
T r = -----dTg
Sr = M 2
(10-4)
Subsonic Regime:
Re
t 1 = Re + S r 4.33 + -------------------------------- e
1 + 0.353T r
(10-5)
t 21 = 0.03Re + 0.48 Re
(10-6)
M
-------------
2 Re
4.5 + 0.38t 21
2
8
t 2 = ------------------------------ + 0.1M + 0.2M e
1 + t 21
(10-7)
------
Re
t 3 = 0.6S r 1 e
(10-8)
24
= ------ + t 2 + t 3
t1
(10-9)
CD
sub
Supersonic Regime:
t 1 = 1.86 M Re
(10-10)
10-5
Droplet Evaporation
1.058 T
2
1
t 2 = 0.9 + t 1 2 + ------- + -----------------------r -------
2
4
Sr
S r
S r
CD
super
t2
= -----------1 + t1
(10-11)
(10-12)
18f
d
p
------ = ------------ ( U ) ------- + g
2
dt
d
d d
(10-13)
dr
----- =
dt
(10-14)
Droplet
Evaporation
m e = 2dm Sh ln ( 1 + Bm )
(10-15)
where Sh is the Sherwood number, m is the mass diffusion coefficient for the gas, and Bm is known as the Spalding number. Let
and be the vapor mass fractions at the droplet surface and in the
10-6
Droplet Evaporation
M d p sat
= ------- --------M p
(10-16)
B m = ------------------1
Sh = 1 + 0.3Re
0.5
Sc
(10-17)
0.333
(10-18)
where Sc is the Schmidt number. Since md = dd3 / 6, the mass conservation equation for the droplet is rewritten in terms of its diameter
4m Sh l n ( 1 + Bm )
d(d )
----------- = --------------------------------------------------dt
d d
(10-19)
dTd
2
md Cd --------- = d q m e L
dt
(10-20)
2K ( T g T d )Nu ln ( 1 + B m )
q = ------------------------------------------------------------------dB m
(10-21)
10-7
Nu = 1 + 0.3Re
0.5
Pr
0.333
(10-22)
dT
Tg Td QL
--------d- = ----------------- ------dt
(10-23)
LShDBm
Q L = ------------------------NuK
(10-24)
where
d d Cd Bm
= --------------------------------------------------12KNu l n ( 1 + Bm )
Droplet Source/
Sink Terms
(10-25)
10-8
Mass Conservation
The net evaporation rate between the beginning and end of a time
step is given as
3
d start d end
m = n d ----------------------------------6
(10-26)
The subscripts start and end refer to the beginning and end of an
integration time step. The number density in a given parcel, represented by n , is calculated from the mass flow rate condition. The
source term for the gas-phase continuity equation is given by
m cell
S m = ----------
(10-27)
where is the volume of the cell in which the mass addition takes
place and m is the accumulated mass source term for the cell. This
term appears on the RHS of equation 10 in addition to any other mass
sources.
If the evaporated droplet mass contributes to a species for which a
mass fraction equation is being solved. then it also appears as a source
term in the mass fraction equation
m i
c ell
S = ----------i
(10-28)
Momentum
Conservation
M = n [ f t ( v U ) + m v ]
(10-29)
10-9
Note that there are two contributions to the momentum source term:
momentum transfer due to drag and momentum of the evaporated
mass. The source term for the gas-phase momentum conservation
equation is given by
M j
c ell
S M = -----------j
(10-30)
Energy
Conservation
E = n [ m h o h ( T g T d ) ]
(10-31)
where h is the total enthalpy of the vapor at the vaporization temperature and h is the heat transfer coefficient calculated from equation
10-21. The source term for the gas-phase enthalpy equation is given
by
E cell
S h = ----------
(10-32)
This term appears on the right hand side of either of the equation 2-5
or equation 2-8 in addition to any other terms such as radiative source
terms.
The source term calculated at the end of an integration time step is
distributed to the neighboring computational cells using a set of
weighting factors. The weighting factor is inversely proportional to
the distance between the droplet location and the cell center. For 2-D
mesh, the source term is distributed to four neighboring cells whereas
for 3-D mesh, it is distributed to the eight neighboring cells.
10-10
Droplet
Dispersion Model
u = ( 2k 3 ) erf ( R )
(10-33)
t int = min ( te , tr )
(10-34)
le = C
32
k
----------
(10-35)
10-11
le
t e = -----u
(10-36)
t r = ln
le
1.0 -------------------
U
(10-37)
4
d
= --- d ----------------------------------( C D U )
3
(10-38)
e
In the case when ------------------- > 1, the above equation has no solu-
tion. This implies that the droplet is captured by the eddy, in which
case tint = te.
Droplet Break-up
Model
10-12
(10-39)
where g and d are the gas and liquid densities, |U-| is the relative
velocity between the gas and the droplet, r is the droplet radius, is
the liquid-gas surface d is the viscosity of the liquid, and y is the
non-dimensional oscillation parameter. Breakup will occur if and
only if y > 1. By linearizing the above equation, an expression for
oscillation can be obtained as,
t
We
We
y ( t ) = -------- + exp ---- [ ( y ( 0 ) + -------- cos t
td
12
12
We
y ( 0 ) -------1
12
+ ---- ( y ( 0 ) + ------------------------- sin t ]
td
(10-40)
d u r
We = -------------------------
(10-41)
2
t d = --------d5 d
(10-42)
10-13
1
= 8 ------------ ------2
3
d r
td
(10-43)
The above parameters are evaluated first to check for the occurrence
of breakup at several stages. In the first stage, if 2 < 0, then the
oscillation of the droplet is negligible and breakup is unlikely to
occur. In that case, the parameters y and y are set to zero. For these
calculations, two permanent arrays are used to store the values of displacement of the drop (y) and time rate of change of displacement
( y ) . If 2> 0, then the amplitude of the undamped oscillation is calculated in the next stage. The amplitude is given by,
2
y 2
We 2
A = y -------- + ----
12
(10-44)
In the above expression, if (We/12+A) < 1.0, then the value of y will
never exceed 1.0 and so break-up is not possible. In this case, the
value of y and y are updated using the following expressions
10-14
n+1
t
We
n
We
= -------- + exp ----- [ ( y + -------- cos
td
12
12
n We
y -------12
1 n
+ ---- ( y + ------------------- sin t ]
td
(10-45)
n + 1
We
-------- y
12
n+1
y
= -------------------------------td
n We
y --------
12
n
t
+ exp ----- [ ( y + ------------------- cos t
t
t
d
d
(10-46)
n We
+ ( y -------- sin t ]
12
In the third stage, break-up is possible in the current time step if (We/
12+A) < 1.0. In this case, the break-up time is calculated as the
smallest root of the following equation
n
We
-------- + A cos [ ( t t ) + ] = 1
12
(10-47)
where
n We
y -------n
12
y
cos = ------------------- and sin = -------A
A
(10-48)
If the break-up time is less than time tn+1, no break-up occurs in the
current time step and hence the y and y values are updated. On the
other hand, if breakup time is greater than tn+1 then break-up occurs
in the current time step. In this case the Sauter mean diameter of the
product drop is calculated as
r
r 32 = ----------------------------------3
d r 2
7 1--- -----------y
--- +
3 8
(10-49)
10-15
Droplet Collision/Coalescence
Droplet Collision/
Coalescence
Model
Nc
2
= ------ ( r c + r d ) d c
(10-50)
Pn = e
n
-------n!
(10-51)
10-16
n = t
(10-52)
b =
R2 + ( r c + rd )
(10-53)
()
We }
(10-54)
where
3
f ( ) = 2.4 + 2.7
r
= ----crd
d d c rd
We = ------------------------------
where is the surface tension between the droplet and gas and We is
the Weber number. If the calculated impact parameter b is less than
the critical parameter, the coalescence occurs, else grazing collision
occurs.
If the outcome of collision is coalescence, then the number of coalescence n for each collector is determined by finding the value of n for
which
10-17
Droplet Collision/Coalescence
n1
k=0
Pk R1 <
Pk
(10-55)
k=1
Then, for each collector drop, n droplets are subtracted from their
associated parcel, and the size, velocity and temperature are appropriately modified. If the outcome of collision is grazing, then the number of droplets involved in the collision is taken to be the minimum
of Nc and Nd. The size, velocity and temperature of the pair of droplets are adjusted such that mass, momentum and energy are conserved.
10-18
11
Radiation Model
Introduction
Blackbody
Radiation
Blackbody is a perfect emitter and absorber of radiation. Using quantum mechanical arguments, it has been shown by Planck and later
verified by experiments that the spectral distribution of emissive
power of a blackbody is given by
2C 1
E = ---------------------------------------C
b
5
2 ( T )
(e
1)
(11-1)
where
11-1
Blackbody Radiation
2
hc
C 1 = hc ;C 2 = -----k
(11-2)
where
h
K
=
=
C2
max T = ---------------------------------------------- C 2 ( max T )
5(1 e
)
(11-3)
E b = T
(11-4)
11-2
Blackbody Radiation
11-3
Radiation Properties
band can be obtained analytically using a series approximation developed by Chang and Rhee, i.e.,
4
T
I b = F ---------
F = F ( 0 2 T ) F ( 0 1 T )
1
e b d
(11-5)
0
F ( 0 1 T ) = ------------------
e b d
0
15
F ( 0 T ) = -----4
n=1
n
2
3 3
e
b b
----------- + --------- + ------ + -----
2
3
n
n
n
n
where
= C 2 T
C 2 = 14388K
Radiation
Properties
(11-6)
E = T
(11-7)
11-4
Radiation Properties
E = T
(11-8)
1
= ---------- e d
4
b
T 0
(11-9)
++ = 1
(11-10)
+ = 1
(11-11)
(11-12)
+ = 1
(11-13)
11-5
Radiation
Characteristics of
Gases
Radiation Characteristics of
1
= ------- ln ( 1 g )
Lm
(11-14)
where Lm is the mean beam length which may be calculated (for optically thin gas radiating to its entire boundary) as
V
L m = 4 --A
(11-15)
Radiative
Transfer
Equation (RTE)
11-6
The integro-differential radiative heat transfer equation for an emitting-absorbing and scattering gray medium can be written as
( )I ( r , ) = ( + )I ( r, ) + I b ( r )
(11-16)
+ ------ I ( r, ) ( ) d
4 = 4
where is the direction of propagation of the radiation beam, I is the
radiation intensity which is a function of both position (r) and direction (), and are the absorption and scattering coefficients respectively, Ib is the intensity of black body radiation at the temperature of
the medium and is the phase function of the energy transfer from
the incoming direction to the outgoing direction . The term on
the left hand side represents the gradient of the intensity in the specified direction . The three terms on the right hand side represent the
changes in intensity due to absorption and out-scattering, emission
and in-scattering respectively. The heat transfer is schematically
shown in figure 11-2.
11-7
I ( r, ) = I b ( r ) + --
n I ( r , ) d
(11-17)
n<0
where I is the intensity of radiant energy leaving a surface at a boundary location, is the surface emissivity, is the surface reflectivity,
and n is the unit normal vector at the boundary location.
11-8
Discrete Ordinate
Method
A number of numerical techniques are available for solving the radiative transfer equation. After a comprehensive review of all the methods, the Discrete-Ordinate Method (Fiveland, 1984) has been
implemented in CFD-ACE because it is well suited for accurately predicting radiative heat transfer for most engineering applications.
Among the several advantages of this method are:
the discrete-ordinate method can be applied to complex geometries even for structured computational grids
the evaluation of the in-scattering term is relatively simple.
the boundary conditions can be imposed more accurately except
for specular reflection from arbitrarily shaped surfaces.
the degree of accuracy of the prediction can be increased by
choosing more ordinates and refining the spatial grid.
the method is aptly suited for finite-volume CFD codes.
I m
I m
I m
x
y
z
+ F I b ;
(11-18)
+ ------ m W m mm I m ;
4
m = 1, M
For the gray model, the subscript should be dropped from the above
equation and F becomes unity. In the nongray model, the radiative
properties are assumed to be functions of wavelength only. For
strongly participating media such as combustion products, in addition
to the wavelength dependence, the radiative properties are functions
of local temperature, pressure, and composition of the gas. Hence,
the radiative properties need to be calculated using either narrowCFD-ACE Theory Manual
11-9
I m = FI b + ------ w m m I m
(11-19)
Im = Im ;
m = m
(11-20)
FT i
I m = ----------------------
(11-21)
Direction
Number
Direction Cosines
Weight
Wm
-0.908248
-0.295876
0.295876
/3
-0.908248
0.295876
0.298576
p/3
-0.295876
-0.908248
0.298576
p/3
-0.295876
-0.295876
0.908248
p/3
-0.295876
0.295876
0.908248
p/3
-0.295876
0.908248
0.298576
p/3
0.295876
-0.908248
0.908248
p/3
0.295876
-0.295876
0.908248
p/3
0.295876
0.295876
0.908248
p/3
10
0.295876
0.908248
0.298576
p/3
11
0.908248
-0.295876
0.295876
p/3
12
0.908248
0.295876
0.298576
p/3
gm
It must be pointed out that the above ordinate set is symmetric only
about the x, y, and z coordinate axes. Hence the specular reflection
boundary condition (used for symmetry) is accurately imposed only
for boundaries that are aligned with the coordinate axes.
11-11
The in-scattering term on the right hand side of equation 11-18 contains the phase function which is generally dependent on the
medium. In CFD-ACE, the medium is assumed to be linearly anisotropic for which the phase function may be written as,
mm = 1 + a o [ m m + m m + m m ]
(11-22)
where a o is an asymmetry factor that lies between -1 and 1. The values -1, 0, 1 denote backward, isotropic and forward scattering, respectively. In CFD-ACE, the in-scattering term is evaluated explicitly
using the previous iteration values and hence the discrete-ordinate
equations are decoupled and the equations are solved sequentially.
Equation 11-18 is numerically integrated over each control volume of
the flow domain for each ordinate direction m using techniques similar to the ones discussed in chapter 4. (See appendix C for details.)
Under conditions of local radiative equilibrium, the net radiative heat
source in a computational cell is the difference between the energy
absorbed and the energy emitted, given by
N band
q net =
=1
I m mm w m 4I b
m m
(11-23)
where is the volume of the cell. This source term is added to the
discretized fluid enthalpy equation. Note that this source term will be
zero for a non-participating medium (k = 0).
Conjugate Heat
Transfer
K g T g = K s T s + Q rad
11-12
(11-24)
AAQ Model
C s T s + C g T g + Q rad
T i = -----------------------------------------------C s + Cg
(11-25)
Ks
Kg
C s = -------- ; C g = -------y s
y g
(11-26)
where
and Qrad is the net radiative heat flux at the interface which is the difference between the radiative energy absorbed and the energy emitted
at the interface. The emission term in the above equation is linearized
to obtain a semi-implicit solution for the interface temperature. The
net gain/loss of heat due to absorption/emission is added as a source
term to the energy equation on both sides of the interface as
C s Q rad
C g Q rad
S s = ------------------- ; S g = -----------------Cs + Cg
Cs + Cg
(11-27)
AAQ Model
CFD-ACE Theory Manual
11-13
AAQ Model
11-14
References
1.
Amsdem. A.A. (1985), KIVA: A Computer Program for Twoand Three-Dimensional Fluid Flows with Chemical Reaction and
Fuel Sprays, Los Alamos National Laboratory Report LA10245-MS.
2.
3.
4.
5.
Crowe, C.T. Sharma, M.P., and Stock, D.E., The ParticleSource-in-Cell (PSI-CELL) Model for Gas-Droplet Flows, J.
Fluids, Engineering, pp. 325-332, 1977.
6.
7.
8.
9.
References-1
References
References
23. Rhie, C.M., and Chow, W.L., (1983), Numerical Study of the
Turbulent Flow Past an Airfoil with Trailing Edge Separation,
AIAA Journal, volume 21, pp. 1525-1532.
24. Rogers, R.C. and Chinitz, W., (1983), AIAA Journal, volume 20,
pp. 586-592.
25. Settari, A. and Aziz, K. (1973), A Generalization of the Additive Correction Methods for the Iterative Solution of Matrix
Equations, SIAM Journal of Numerical Analysis, volume 10, pp.
506, 1973.
26. Thompson, J.F., Warsi, Z.U.A., and Mastin, C.W., (1985),
Numerical Grid Generation, North-Holland, New York.
27. Van Doormaal, J.P. and Raithby, G.D. (1984), Enhancements of
the SIMPLE Method for Predicting Incompressible Fluid Flows,
Numerical Heat Transfer, volume 7, pp. 147-163.
28. Westbrook, C.K., and Dryer, F.L., Combined Science and Technology, volume 27, pp. 31-43, 1981.
29. Wilcox, D.C., (1991), A Half Century Historical Review of the
k- Model, AIAA-91-0615.
30. Yang, H.Q., Habchi, S.D., and Przekwas, A.J., (1992), A General Strong Conservation Formulation of Navier-Stokes Equations in Non-Orthogonal Curvilinear Coordinates, AIAA-920187.
31. Yakhot, V. and Orszag, S.A., (1986), Renormalization Group
Analysis of TurbulenceI: Basic Theory, J. Scientific Computing, volume 1, pp. 151.
32. Yahkot, V., Orszag, S.A., Thangam, S., Gatski, T.B., and Speziale, C.G., (1992), Development of Turbulence Models for
Shear Flows by a Double-Expansion Technique, Physics Fluids,
volume 4, pp. 15101520.
References-3
References
References-4
Appendix A
Introduction
Fluid Properties
Density
Constant Density
This law generally applies to gaseous flows from subsonic to hypersonic range.
A-1
Viscosity
pM
= -------RT
(A-1)
Inverse Temperature This is a simplified version of the perfect gas law. In low-Mach-number gaseous flows involving reaction or heat transfer, the density varModel
ies primarily due to temperature variation. The density is obtained
from the perfect gas law by replacing p with pref a reference pressure
specified by the user. In general, pref equals the average static pressure in the flow domain.
Viscosity
Constant Viscosity
This model generally applies to isothermal liquid and low-Machnumber gaseous flows without any chemical reaction or heat transfer.
Power Law
T
= 0 ------
T
(A-2)
Sutherlands Law
AT
= ------------B+T
A-2
(A-3)
Specific Heat
Here, the constants A and B depend on the gaseous species and should
be specified by the user.
Specific Heat
Constant Specific
Heat
Thermal
Diffusivity
= --
(A-4)
The thermal conductivity K is obtained from the diffusivity and specific heat.
K = Cp
(A-5)
t = -----t
t
(A-6)
Mass Diffusivity
The mass diffusion coefficient, a fluid property that appears in mixture fraction and species mass fraction equations, is obtained from
relations identical to equation A-4 and equation A-6 but with and t
representing molecular and turbulent Schmidt numbers.
A-3
A-4
Mass Diffusivity
Appendix B
Introduction
JANNAF Tables
The Joint Army, Navy, NASA, and Air Force (JANNAF) Tables are a
set of thermodynamic correlations assembled for use by the U.S. Military and NASA organizations. These correlations can be used to estimate the specific heat, enthalpy and entropy of a number of species
that are encountered in combustors, reactors, and various kinds of
flow equipment. The basic methodology of JANNAF Tables is
briefly presented below.
In JANNAF methodology, the thermodynamic quantities, viz. specific
heat, enthalpy and entropy, are evaluated at the standard atmospheric
pressure (po = 1.01325 x 105 N/m2). The dependence of these quantities on temperature is expressed in terms of polynomials. Several
coefficients are needed for each of the two temperature ranges, 300 to
1000 K and 1000 to 5000 K. Thus, the JANNAF Tables contain a set
of 14 coefficients for each species.
The molar specific heat of a species i is expressed in a nondimensional form as
C p
2
3
4
-------i = z 1 + z 2 T + z 3 T + z 4 T + z 5 T
R
(B-1)
where R is the universal gas constant and T is the local fluid temperature expressed in absolute scale. For example, in SI units, T is
B-1
Introduction
h i
C p = -------
i
T p
(B-2)
z2
z3 2
h
------i- = z 1 + ----- T + ----- T +
2
3
RT
z 3 z 4 z
----4- T + ----5- T + ----64
5
T
(B-3)
Note that the species enthalpy in equation B-1 includes the enthalpy
of formation and the sensible enthalpy.
h i = h f
298
Cp dT
i
(B-4)
298
h f
298
dh Tfp
ds i = -------i -------T
p
(B-5)
Introduction
p
s i = s i R ln -----
p
(B-6)
si
----- = z 1 ln T + z 2 T +
R
z 2
----3- T +
2
z 3 z 4
----4- T + ----5- T + z 7
3
4
(B-7)
s i = s
298
C
p
-------i dT
T
(B-8)
298
298
reference quantities.
In summary, equation B-1, equation B-3 and equation B-7 are used to
obtain molar quantities at standard atmospheric pressure. Although
the JANNAF coefficients are tabulated for a single pressure, little
error is incurred in using these tables over the range of pressures normally encountered in combustion problems. This is because the specific heat and enthalpy of gases happen to be weak functions of
pressure.
The species molar quantities are converted to per unit mass quantities by dividing with the species molecular weight.
M
Cp = C
pi
i
i
h i = h i M i
(B-9)
(B-10)
B-3
Introduction
s i = s i M i
(B-11)
Thermodynamic quantities for the mixture are obtained from individual species quantities and mass fractions.
C p = i C p
i
(B-12)
h = i h i
(B-13)
s = i s i
(B-14)
The JANNAF correlations are used mainly for two purposes in CFDACE.
1.
hh
e. T = T + --------------
Cp
B-4
Appendix C
Introduction
Discretization of Discrete
Ordinate Equations
1 J
1 J
1 J
--- ------ ----- m I m + --- ------ ----- m I m + --- ------ ----- I m
J h 1
J h 2
J h 3
= ( + )I m + I b + ------ Wm mm I m
4
(C-1)
m = m F 1x + m F 1y + m F 1z
(C-2)
m = m F 2x + m F 2y + m F 2z
(C-3)
C-1
Introduction
m = m F 3x + m F 3y + m F 3z
(C-4)
F1x, F1y, etc. are defined in equation 89. Equation C-1 can be integrated over the control volume, shown in figure 11-1, to obtain
A1 I A 1 I
me w
m w
m e e mw
2
2
+ mn An I m m s As I m
n
s
(C-5)
3
3
+ m h Ah I m ml Al I m
h
l
= ( + ) I m + I b + ------ W m mm I m
p
p
p
4
,A
, and A
represent the
C-2
I m ( 1 d )I m
p
w
I m = -------------------------------------e
d
(C-6)
I m ( 1 d )I m
p
-s
I m = ------------------------------------n
d
(C-7)
Introduction
Im ( 1 d )Im
p
Im = -------------------------------------t
k
d
(C-8)
BwIm + BsIm + Bl Im + S1 S2
w
s
t
I m = -----------------------------------------------------------------------------p
Bp
(C-9)
where
1
1
B w = dA w mw + ( 1 d )A e me
2
2
B s = dA s m s + ( 1 d )An m n
3
3
B l = dA l m l + ( 1 d )A h m k
1
2
3
B p = A w m e + A w mn + A h m k + d ( + )
S 1 = dk I b
S 2 = d ------ Wm mm I m
p
4
Equation C-9 is appropriate when all three direction cosines are positive and the direction of integration proceeds in a direction of increasing , and . If a direction cosine is negative, the corresponding
direction of integration is reversed and the integration sweep is started
from the appropriate corner of the domain. Thus, for each set of
direction cosines, the intensities at all the cell centers are obtained by
marching in appropriate directions. This procedure is repeated for all
the ordinate directions. The in-scattering term, i.e., the last term on
C-3
Introduction
C-4
2.
3.
4.
5.
6.
oYes
7.
8.
9.
As compared to the 8.5 x 11" book size, is the new book size preferable?
Yes
No
Not sure
o
o
o
o
o
o
o
o
o
oNo
o
o
o
o
o
o
o
11. Please enter here any comments you have about the documentation:
___________________________________________________________________________
___________________________________________________________________________
___________________________________________________________________________
Thank you for taking the time to complete this survey. Please send this information to the following:
FAX: (256) 726-4806 or fold and seal and return to CFD Research Corporation, 215 Wynn
Drive Suite 501, Huntsville, AL 35805
From:
Place
Stamp
Here