Computational Fluid Dynamics

It is usually abbreviated as CFD, is a branch of fluid mechanics that

uses numerical methods and algorithms to solve and analyze problems that
involve fluid flows. Computers are used to perform the calculations required
to simulate the interaction of liquids and gases with surfaces defined
by boundary conditions. With high-speed supercomputers, better solutions
can be achieved. Ongoing research yields software that improves the
accuracy and speed of complex simulation scenarios such as transonic
or turbulent flows. Initial validation of such software is performed using
a wind tunnel with the final validation coming in full-scale testing, e.g. flight
tests.

The fundamental basis of almost all CFD problems are the NavierStokes
equations, which define any single-phase fluid flow. These equations can
be simplified by removing terms describing viscosity to yield the Euler
equations. Further simplification, by removing terms describing vorticity
yields the full potential equations. Finally, these equations can be linearized
to yield the linearized potential equations.

Methodology :-

In all of these approaches the same basic procedure is

followed.During preprocessing
The geometry (physical bounds) of the problem is defined.
The volume occupied by the fluid is divided into discrete cells
(the mesh). The mesh may be uniform or non- uniform.
The physical modeling is defined for example, the equations of
motions + enthalpy + radiation + species conservation.
Boundary conditions are defined. This involves specifying the
fluid behavior and properties at the boundaries of the problem.
For transient problems, the initial conditions are also defined.
The simulation is started and the equations are solved iteratively
as a steady-state or transient.
Finally a postprocessor is used for the analysis and visualization
of the resulting solution.
Discretization methods :-
The stability of the chosen discretization is generally established
numerically rather than analytically as with simple linear problems. Special
care must also be taken to ensure that the discretization handles
discontinuous solutions gracefully. The Euler equations and NavierStokes
equations both admit shocks, and contact surfaces.
Some of the discretization methods being used are :-

1. Finite volume method

The finite volume method (FVM) is a common approach used in CFD. The
governing equations are solved over discrete control volumes. Finite
volume methods recast the governing partial differential equations (typically
the Navier-Stokes equations) in a conservative form, and
then discretize the new equation. This guarantees the conservation of
fluxes through a particular control volume. The finite volume equation yields
governing equations in the form,

where is the vector of conserved variables, is the vector of fluxes

(see Euler equations or NavierStokes equations), is the volume of
the control volume element, and is the surface area of the control
volume element.

2. Finite element method :-

The finite element method (FEM) is used in structural analysis of solids,

but is also applicable to fluids. However, the FEM formulation requires
special care to ensure a conservative solution. The FEM formulation has
been adapted for use with fluid dynamics governing equations.Although
FEM must be carefully formulated to be conservative, it is much more
stable than the finite volume approach. However, FEM can require more
memory than FVM.
In this method, a weighted residual equation is formed:
 where is the equation residual at an element vertex , is the
conservation equation expressed on an element basis, is the
weight factor, and is the volume of the element.

3. Finite difference method :-

The finite difference method (FDM) has historical importance and is
simple to program. It is currently only used in few specialized codes.
Modern finite difference codes make use of an embedded boundary
for handling complex geometries, making these codes highly efficient
and accurate. Other ways to handle geometries include use of
overlapping grids, where the solution is interpolated across each grid.

where is the vector of conserved variables, and , , and

are the fluxes in the , , and directions respectively.

Navier-Stokes Theorem :-
A simplification of the resulting flow equations is obtained when
considering an incompressible flow of a Newtonian fluid. The
assumption of incompressibility rules out the possibility
of sound or shock waves to occur; so this simplification is invalid if
these phenomena are important. The incompressible flow assumption
typically holds well even when dealing with a "compressible" fluid
such as air at room temperature at low Mach numbers (even when
flowing up to about Mach 0.3). Taking the incompressible flow
assumption into account and assuming constant viscosity, the
NavierStokes equations will read, in vector form:

NavierStokes equations (Incompressible flow)

Here f represents "other" body forces (forces per unit volume), such
as gravity or centrifugal force. The shear stress term becomes the
useful quantity ( is the vector Laplacian) when the fluid is
assumed incompressible, homogeneous and Newtonian, where is the
(constant) dynamic viscosity.
It's well worth observing the meaning of each term (compare to
the Cauchy momentum equation):

Note that only the convective terms are nonlinear for incompressible
Newtonian flow. The convective acceleration is an acceleration
caused by a (possibly steady) change in velocity over position, for
example the speeding up of fluid entering a converging nozzle.
Though individual fluid particles are being accelerated and thus are
under unsteady motion, the flow field (a velocity distribution) will not
necessarily be time dependent.
Another important observation is that the viscosity is represented by
the vector Laplacian of the velocity field (interpreted here as the
difference between the velocity at a point and the mean velocity in a
small volume around). This implies that Newtonian viscosity
is diffusion of momentum, this works in much the same way as
the diffusion of heat seen in the heat equation (which also involves
the Laplacian).
If temperature effects are also neglected, the only "other" equation
(apart from initial/boundary conditions) needed is the mass continuity
equation. Under the incompressible assumption, density is a
constant and it follows that the equation will simplify to:

This is more specifically a statement of the conservation of

volume (see divergence).
These equations are commonly used in 3 coordinates
systems: Cartesian, cylindrical, and spherical. While the Cartesian
equations seem to follow directly from the vector equation above,
the vector form of the NavierStokes equation involves
some tensor calculus which means that writing it in other
coordinate systems is not as simple as doing so for scalar
equations (such as the heat equation).
 Air Foil
Studying the flow over an airfoil, particularly for a
prototype,iscritical in the aerodynamic design ofaircrafts.
The shape ofthe airfoil, the boundary conditions,
andthedragandliftforcesgenerated on theairfoil
areimportantin identifying the performance
characteristicsofthe airfoilduring flight. Similar
principlescanbeapplied tohydrofoils(artificially-
madeinboatsornaturally- foundinaquatic animals).

Problem Solving Techniques :-

VisualizecasewithParaView.
Setboundaryconditions andsolverparameters.
Run withOpenFOAMsolverfor=0o andanalyze.
Rotatemeshandsolveforother anglesofattack.
Examine thedragandliftforcecoefficientsvs..
Whichisthecritical angleofattack forthisairfoil??

Figure 1:Angleofattack andforcesacting

onanairfoil

1.2 Forcesactingon anairfoil

Drag:inthedirection ofthefreestream velocity(Fig.1).
Lift: perpendiculartothedirection ofthefreestream
velocity(Fig.1).
 Thrust:forcegeneratedthroughthe reactionofaccelerating
amassoffluid, intheoppositedirection fromtheacceleratedfluid.
Weight:theforceonanobjectduetogravity.

1.3 Terminology
Airfoil:thecross-sectional shapeofanaircraft wing.

(Figure -1)

Problem Specification :-Study Of Flow Of Air Over An Airfoil.

 Step 1: Create Geometry in GAMBIT.

In an external flow such as that over an airfoil, we have to define a farfield

boundary and mesh the region between the airfoil geometry and the farfield
boundary. It is a good idea to place the farfield boundary well away from
the airfoil since we'll use the ambient conditions to define the boundary
conditions at the farfield. The farther we are from the airfoil, the less effect it
has on the flow and so more accurate is the farfield boundary condition.

1.1 Start GAMBIT

Create a new directory called airfoil and start GAMBIT from that directory
by typing gambit -id airfoil at the command prompt.

Under Main Menu, select Solver > FLUENT 5/6 since the mesh to be
created is to be used in FLUENT 6.0.

1.2 Airfoil Geometry

To specify the airfoil geometry, we'll create a list of vertices along the
surface and have GAMBIT join these vertices to create two edges,
corresponding to the upper and lower surfaces of the airfoil. We'll then split
these edges into 4 distinct edges to help us control the mesh size at the
surface.

1.3 Create Farfield Boundary

Next, we created the following farfield boundary.

Farfield boundary is created by creating vertices and joining them

appropriately to form edges.

Operation Toolpad> Geometry Command Button > Vertex Command

Button > Create Vertex

Create the following vertices by entering the coordinates under Global and
the label under Label:

Label X y z
A 1 12.5 0
B 21 12.5 0
C 21 0 0
D 21 -12.5 0
E 1 -12.5 0
F -11.5 0 0
G C 0 0

Now we can create the edges using the vertices created.

Operation Toolpad> Geometry Command Button > Edge Command

Button > Create Edge

Create the edge AB by selecting the vertex A followed by vertex B. Enter

AB for Label. Click Apply. GAMBIT will create the edge. You will see a
message saying something like "Created edge: AB'' in
the Transcript window.

Similarly, create the edges BC, CD, DE, EG, GA and CG.

Next we'll create the circular arc AF. Right-click on the Create Edge button
and select Arc.

In the Create Real Circular Arc menu, the box next to Center will be yellow.
That means that the vertex selected will be taken as the center of the arc.
Select vertex G and click Apply. Now the box next to End Points will be
highlighted in yellow. Select vertex A and then vertex F. Enter AF
under Label. Click Apply.

Similarly, create an edge corresponding to arc EF.

1.4 Create Faces

The edges we have created can be joined together to form faces. We will
need to define three faces as shown in the image above. Two rectangular
faces, rect1 and rect2 lie to the right of the airfoil. The third face, circ1
consists of the area outside of the airfoil but inside of the semi-circular
boundary.

Operation Toolpad> Geometry Command Button > Face Command

Button > Form Face

This brings up the Create Face From Wireframe menu. Recall that we had
selected vertices in order to create edges. Similarly, we will select edges in
order to form a face.

To create the face rect1, select the edges AB, BC, CG, and GA. Enter
rect1for the label and click Apply. GAMBIT will tell that it has "Created
face: rect1'' in the transcript window.

Similarly, create the face rect2 by selecting ED, DC, CG and GE.

To create the last face we will need to make two seperate faces, one for
the outer boundary and one for the airfoil and then subtract the airfoil from
the boundary . Create semi-circular face circ1 by selecting GA, AF, FE and
EG and enter circ1 for the label. Create the face for the airfoil by selecting
corresponding edges. Subtract the airfoil from circ1.

Operation Toolpad> Geometry Command Button > Face Command

Button

Right click on the Boolean Operations Buttonand select Subtract

The Face box will be highlighted yellow. Shift click to select circ1, the outer
semi-circular boundary. Then select the lower box labeled Subtract
Faces which will allow you to select faces to subtract from our outer
boundary. Select the airfoil face and click appl

Step 2: Mesh Geometry in GAMBIT

2.1 Mesh Faces

We'll mesh each of the 3 faces separately to get our final mesh. Before we
mesh a face, we need to define the point distribution for each of the edges
that form the face i.e. we first have to mesh the edges. We'll select the
mesh stretching parameters and number of divisions for each edge based
on three criteria:

1. We'd like to cluster points near the airfoil since this is where the flow
is modified the most; the mesh resolution as we approach the farfield
boundaries can become progressively coarser since the flow
gradients approach zero.
2. Close to the surface, we need the most resolution near the leading
and trailing edges since these are critical areas with the steepest
gradients.
3. We want transitions in mesh size to be smooth; large, discontinuous
changes in the mesh size significantly decrease the numerical
accuracy.

The edge mesh parameters we'll use for controlling the stretching
are successive ratio, first length and last length. Each edge has a direction
as indicated by the arrow in the graphics window. The successive ratio R is
the ratio of the length of any two successive divisions in the arrow direction
as shown below.

Operation Toolpad> Mesh Command Button > Edge Command

Button > Mesh Edges

Select the edge GA. The edge will change color and an arrow and several
circles will appear on the edge. The arrow is pointing upwards. We can
reverse the direction of the edge by clicking on the Reverse button in
the Mesh Edges menu. Enter a ratio of 1.15. This means that each
successive mesh division will be 1.15 times bigger in the direction of the
arrow. Select Interval Count under Spacing. Enter 45 forInterval Count.
Click Apply. GAMBIT will create 45 intervals on this edge with a
successive ratio of 1.15.

For edges AB and CG, we'll set the First Length (i.e. the length of the
division at the start of the edge) rather than the Successive Ratio. Repeat
the same steps for edges BC, AB and CG with the following specifications:

Arrow Successive Interval

Edges
Direction Ratio Count
GA
and Upwards 1.15 45
BC

Arrow First Interval

Edges
Direction Length Count
AB
Left to
and 0.02c 60
Right
CG

Note that later we'll select the length at the trailing edge to be 0.02c so that
the mesh length is continuous between IG and CG, and HG and CG.

Now that the appropriate edge meshes have been specified, mesh the face
rect1:

Operation Toolpad> Mesh Command Button > Face Command

Button > Mesh Faces

Select the face rect1. The face will change color. We can use the defaults
of Quad (i.e. quadrilaterals) and Map. Click Apply.
 The meshed face look as follows:

Next mesh face rect2 in a similar fashion. The following table shows the
parameters to use for the different edges:

Arrow Successive Interval

Edges
Direction Ratio Count
EG
and Downwards 1.15 45
CD

Arrow First Interval

Edges
Direction Length Count
Left to
DE 0.02c 60
Right
 The resultant mesh should be symmetric about CG as shown in the figure
below.

2.2 Split Edges

Next, we will split the top and bottom edges of the airfoil into two edges so
that we have better control of the mesh point distribution. Figure of the
splitting edges is shown below.

We need to do this because a non-uniform grid spacing will be used

for x<0.3c and a uniform grid spacing for x>0.3c. To split the top edge into
HI and IG, select

Operation Toolpad> Geometry Command Button > Edge Command

Button > Split/Merge Edge

Make sure Point is selected next to Split with in the Split Edge window.
 Select the top edge of the airfoil by Shift-clicking on it. We should see
something similar to the picture below:

We'll use the point at x=0.3c on the upper surface to split this edge into HI
and IG. To do this, enter 0.3 for x: under Global. If c is not equal to one,
enter the value of 0.3*c instead of just 0.3.For instance, if c=4, enter 1.2.

We should see that the white circle has moved to the correct location on
the edge.
Click Apply. We will see a message saying ``Edge edge.1 was split, and
edge edge.3 created'' in the Transcript window.

Note the yellow marker in place of the white circle, indicating the original
edge has been split into two edges with the yellow marker as its dividing
point.
Repeat this procedure for the lower surface to split it into HJ and JG. Use
the point at x=0.3c on the lower surface to split this edge.

Finally, let's mesh the face consisting of circ1 and the airfoil surface. For
edges HI and HJ on the front part of the airfoil surface, use the following
parameters to create edge meshes:

Arrow Last Interval

Edges
Direction Length Count
From H
HI 0.02c 40
to I
From H
HJ 0.02c 40
to J

For edges IG and JG, we'll set the divisions to be uniform and equal to
0.02c. Use Interval Size rather than Interval Count and create the edge
meshes:

Arrow Successive Interval

Edges
Direction Ratio Size
 IG
Left to
and 1 0.02c
Right
JG

For edge AF, the number of divisions needs to be equal to the number of
divisions on the line opposite to it, in this case, the upper surface of the
airfoil(this is a subtle point; chew over it). To determine the number of
divisions that GAMBIT has created on edge IG.

Operation Toolpad> Mesh Command Button > Edge Command

Button >Summarize Edge Mesh

Select edge IG and then Elements under Component and click Apply.
This will give the total number of nodes (i.e. points) and elements (i.e.
divisions) on the edge in the Transcript window. The number of divisions on
edge IG is 36. So the Interval Count for edge AF is NHI+NIG= 40+36= 76.

Similarly, determine the number of divisions on edge JG. This comes out
as 35 for the current geometry. So the Interval Count for edge EF is 75.
Create the mesh for edges AF and EF with the following parameters:

Arrow First Interval

Edges Direction Length Count

From A
AF 0.02c 40+NIG
to F
From E
EF 0.02c 40+NJG
to F
 Figure : Resultant Mesh

Step 3: Specify Boundary Types in GAMBIT

We'll label the boundary AFE as farfield1, ABDE as farfield2 and the airfoil
surface as airfoil. These will be the names that show up under boundary
zones when the mesh is read into FLUENT.

3.1 Group Edges

We'll create groups of edges and then create boundary entities from these
groups.

First, we will group AF and EF together.

Operation Toolpad> Geometry Command Button > Group Command

Button > Create Group

Select Edges and enter farfield1 for Label, which is the name of the group.
Select the edges AF and EF.

Note that GAMBIT adds the edge to the list as it is selected in the GUI.
Click Apply.

Similarly, create the other two farfield groups. You should have created a
total of three groups:

Group Name Edges in Group

farfield1 AF,EF

farfield2 AB,DE

farfield3 BC,CD

Airfoil HI,IG,HJ,JG (name might vary)

3.2 Define Boundary Types

Now that we have grouped each of the edges into the desired groups, we
can assign appropriate boundary types to these groups.

Operation Toolpad> Zones Command Button > Specify Boundary

Types
Select any edge belonging to the airfoil surface and that will select the
airfoil group. Next to Name:, enter airfoil. Leave the Type as WALL.

Click Apply.
 Step 4: Set Up Problem in FLUENT

4.1 Launch FLUENT

Start > Programs > Fluent Inc> FLUENT 6.3.26

Select 2ddp from the list of options and click Run.

4.2 Import File

Main Menu > File > Read > Case...

Navigate to working directory and select the airfoil.msh file. Click OK.

The following should appear in the FLUENT window:

Check that the displayed information is consistent with our expectations of

the airfoil grid.
4.3 Analyze Grid

Grid > Info > Size

How many cells and nodes does the grid have?

Display > Grid

Note what the surfaces farfield1, farfield2, etc. correspond to by selecting

and plotting them in turn.

Zoom into the airfoil.

Where are the nodes clustered? Why?

Define Properties

Define > Models > Solver...

Under the Solver box, select Pressure Based.

Click OK.
Define > Models > Viscous

Select Inviscid under Model.

Click OK.

Define > Models > Energy

The speed of sound under SSL conditions is 340 m/s so that our
freestream Mach number is around 0.15. This is low enough that we'll
assume that the flow is incompressible. So the energy equation can be
turned off.

Make sure there is no check in the box next to Energy Equation and
click OK.

Define > Materials

Make sure air is selected under Fluid Materials.

Set Density to constant and equal to 1.225 kg/m3.
Click Change/Create.

Define > Operating Conditions

We'll work in terms of gauge pressures in this example. So set Operating

Pressure to the ambient value of 101,325 Pa.

Click OK.

Define > Boundary Conditions

Set farfield1 and farfield2 to the velocity-inlet boundary type.

For each, click Set.... Then, choose Components under Velocity

Specification Method and set the x- and y-components to that for the
freestream. For instance, the x-component is 50*cos(1.2)=49.99. (Note that
1.2 is used as our angle of attack instead of 2 to adjust for the error
caused by assuming the airfoil to be 2D instead of 3D.)

Click OK.

Set farfield3 to pressure-outlet boundary type, click Set... and set

the Gauge Pressure at this boundary to 0. Click OK.

Step 5: Solve

Solve > Control > Solution

Take a look at the options available.

Under Discretization,
set Pressure to PRESTO! and Momentum to Second-Order Upwind.
Click OK.

Solve > Initialize > Initialize...

Once again, we'll set these values to be equal to those at the inlet (to
review why we did this look back to the tutorial about CFG programs) .
Select farfield1 under Compute From.

Click Init.

Solve > Monitors > Residual...

Now we will set the residual values (the criteria for a good enough
solution). Once again, we'll set this value to 1e-06.
Click OK.

Solve > Monitors > Force...

Under Coefficient, choose Lift. Under Options, select Print and Plot.
Then, Choose airfoil under Wall Zones.

Lastly, set the Force Vector components for the lift. The lift is the force
perpendicular to the direction of the freestream. So to get the lift coefficient,
set X to -sin(1.2)=-020942 and Y to cos(1.2)=0.9998.

Click Apply for these changes to take effect.

Similarly, set the Force Monitor options for the Drag force. The drag is
defined as the force component in the direction of the freestream. So
under Force Vector, set X to cos(1.2)=0.9998 and Y to
sin(1.2)=0.020942 Turn on only Print for it.

Report > Reference Values

Now, set the reference values to set the base cases for our iteration.
Select farfield1 under Compute From.

Click OK.

Note that the reference pressure is zero, indicating that we are measuring
gage pressure.

Main Menu > File > Write > Case...

Save the case file before you start the iterations.

Solve > Iterate

Make note of your findings, make sure you include data such as;

What does the convergence plot look like?

How many iterations does it take to converge?

How does the Lift coefficient compared with the experimental data?

Main Menu > File > Write > Case & Data...

Save case and data after you have obtained a converged solution.

Step 6: Analyze Results

Plot Velocity Vectors

Let's see the velocity vectors along the airfoil.

Display > Vectors

Use the default setting by clicking Display.

(Figure -Velocity Flow)

As can be seen, the velocity of the upper airfoil is faster than the velocity on
the lower airfoil.
 -

On the leading edge, we see a stagnation point where the velocity of the
flow is nearly zero. The fluid accelerates on the upper surface as can be
seen from the change in colors of the vectors.

On the trailing edge, the flow on the upper surface decelerates and
converge with the flow on the lower surface.

Plot Pressure Coefficient

Pressure Coefficient is a dimensionless parameter defined by the

equation where is the static pressure, is the

reference pressure, and is the reference dynamic pressure defined
by . The reference pressure, density, and velocity are defined in
the Reference Values panel in Step 5.

Plot > XY Plot...

Change the Y Axis Function to Pressure..., followed by Pressure

Coefficient. Then, select airfoil under Surfaces.

Click Plot.

The negative part of the plot is upper surface of the airfoil as the pressure
is lower than the reference pressure.
Plot Pressure Contours

Plot static pressure contours.

Display > Contours...

Select Pressure... and Static Pressure from under Contours Of.

Click Display. Check also the Filled and Draw Grid under Options menu.

From the figure, we see that in one grid, there is no more than 3 different
pressure contours which suggests that our mesh is fine enough.

How can we compare the pressure contour with velocity vector plot? We
see that the pressure on the upper surface is negative while the velocity on
the upper surface is higher than the reference velocity. Whenever there is
high velocity vectors, we have low pressures and vise versa. The
phenomenon that we see comply with the Bernoulli equation.
STUDY OF FLUID FLOW OVER AN AIRFOIL
USING CFD

A Project
Submitted in partial fulfillment of the requirements for
The award of the Degree of
BACHELOR OF ENGINEERING
IN
MECHANICAL ENGINEERING
By

RAJAT TYAGI ( BE/5514/08)

ARVIND PAVITHARAN ( BE/5675/08)

ASHIQUE DHOULIYAN ( BE/5670/08)

PALLAVI PRAKASH ( BE/5595/08)

DEPARTMENT OF MECHANICAL ENGINEERING

BIRLA INSTITUTE OF TECHNOLOGY
MESRA
 DECLARATION CERTIFICATE

This is to certify that the thesis entitled, STUDY OF FLUID FLOW

OVER AN AIRFOIL submitted by Mr. RajatTyagi ,Mr. ArvindPavithran ,
Mr. AshiqueDhouliyanand Ms. PallaviPrakashin partial fulfillment of the
requirements for the award of Bachelor of Engineering Degree in
MECHANICALat Birla Institute of Technology, Mesra, Ranchi is an authentic
work carried out by them under my supervision and guidance. To the best of my
knowledge, the matter embodied in the thesis has not been submitted to any other
University/ Institute for the award of any degree diploma.

Date:

Prof. SoumenMandal

Mechanical Dept.
Birla Institute of Technology
Mesra- 835215

Head

Mechanical Engineering.
Birla Institute of Technology
Mesra-835215
 CERTIFICATE OF APPROVAL

The foregoing thesis entitled Study Of Fluid Flow Over An Air Foil, is
hereby approved as a creditable study of research topic and has been presented in
satisfactory manner to warrant its acceptance as prerequisite to the degree for
which it has been submitted.

It is understood that by this approval, the undersigned do not necessarily

endorse any conclusion drawn or opinion expressed therein, but approve the thesis
for the purpose for which it is submitted.

(Internal Examiner) (External Examiner)

(Chairman)

Head of the Department

 ACKNOWLEDGEMENT

We express our sincere gratitude and indebtedness to the thesis guideProf.

SoumenMandalfor his initiative in this field of research, for hisvaluable guidance,
encouragement and affection for the successful completion of this work. His
sincere sympathies and kind attitude always encouraged us to carry out the present
work firmly. Weexpress our thankfulness to Prof.P.K.Chaudhary, Head of the
Department of Mechanical Engineering, Birla Institute of Technology Patna for
providing us with best facilities in the Department and his timely suggestions. We
would like to thank FluentInc. for providing the simulation package.

Last but not least we would like to thank all our friends and well-wishers
who were involved directly or indirectly in successful completion of the present
work.

Mr. RajatTyagi

Mr. ArvindPavitharan

Mr. AshiqueDhouliyan

Ms. PallaviPrakash
 REFERENCES

www.cornfluence.cornell.edu
www.cfluid.com
www.cfd-online.com
Milne-Thomson, L.M. (1973). Theoretical Aerodynamics. Dover Publications.
Surana, K.A.; Allu, S.; Tenpas, P.W.; Reddy, J.N. (February 2007). "k-version
of finite element method in gas dynamics: higher-order global differentiability
numerical solutions". International Journal for Numerical Methods in
Engineering.
Launder, B. E.; D. B. Spalding (1974). "The Numerical Computation of
Turbulent Flows". Computer Methods in Applied Mechanics and Engineering.
Wilcox, David C. (2006). Turbulence Modeling for CFD (3 ed.). DCW
Industries, Inc..
Pope, S. B. (2000). Turbulent Flows. Cambridge University Press.
Fox, Rodney (2003). Computational models for turbulent reacting flows.
Cambridge University Press.
Anderson, John D. (1995). Computational Fluid Dynamics: The Basics With
Applications. Science/Engineering/Math. McGraw-Hill Science.