R

Exploring OpenFOAM

DR
AF
T

Comflics
Sadiq Huq
Shuaib Ahmed Idrees
Mehran Saeedi Ramakrishna Nanjundiah

Beta Release, August 2014.

www.comflics.de

Disclaimer

DR
AF
T

This offering is not approved or endorsed by OpenCFD Limited, the producer of the OpenFOAM software and owner of the OPENFOAMrand OpenCFDrtrade marks.

ii

Preface
R
The response to OpenFOAM
tutorials in Youtube has been overwhelming. The videos
R
were made to give a quick start to using OpenFOAM
. Continuing our efforts at making
R

OpenFOAM easy to understand, we now bring to you an ebook. Each chapter will demonstrare a key topic in Computational Fluid Dynamics, the simulation cases used in this book
will also be made available to you. We encourage you to tweak the simulations, experiment
new possibilities and whenever possible share your findings in a public domain.

You can download the files used in this ebook from https://github.com/Comflics/comflicsTutor.
And you can also follow our blog, www.comflics.de, for updates.

DR
AF
T

Munich, August 27, 2014.

iii

Contents
1 Introduction to OpenFOAM - (coming soon)
S. Huq

2 Laminar Vortex Shedding

M. Saeedi

2.4
2.5

Fluid Mechanics . . . .
Numerics . . . . . . .
OpenFOAM . . . . . .
2.3.1 Preprocessing .
2.3.2 Solver . . . . .
2.3.3 Post processing
Exercise . . . . . . . .
References . . . . . . .

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

.
.
.
.
.
.
.
.

3
4
7
7
11
14
16
16

DR
AF
T

2.1
2.2
2.3

1 Introduction to OpenFOAM - (coming

soon)

DR
AF
T

S. Huq

2 Laminar Vortex Shedding

M. Saeedi

2.1 Fluid Mechanics

DR
AF
T

Flow around a cylinder is a classic problem in fluid mechanics. It is a problem of practical

importance: Off-shore structures, bridge piers, single silos and to some extent cooling towers
are typical cases where we face flow over cylindrical structures. Vortex shedding is an
important phenomena which occurs in flow around these types of structures for a wide range
of Reynolds numbers. The frequency of vortex shedding and analysis of vortex induced
vibrations play an important role in design of these structures. In this tutorial laminar flow
around a cylinder is simulated using OpenFOAM and the results are validated by comparing
them with experimental data. First of all, physical aspect of flow around a cylinder is reviewed.
The numerical aspects of simulating the case together with case setup in OpenFoam are
discussed, the chapter ends with validating the results by comparing them with experimental
data.

The most decisive non-dimensional parameter for describing the flow pattern around a cylinder
is the Reynolds number:
Re =

UL

(2.1)

It is defined as the ratio of inertial to viscous forces. At low Reynolds numbers viscous
forces dominate the flow and the flow is laminar, as the Reynold number increases viscous
forces cease to dominate the flow and the flow becomes turbulent. For very low Reynolds
numbers Re < 5 there is no flow separation, fluid particles from upstream of the flow can be
divided to two groups flowing symmetrically around upper and lower part of the cylinder
and reaching together the back of the cylinder without separation. As the Reynolds number
is increased a small separation region is formed behind the cylinder where two eddies of
opposite sign are formed (twin vortices), these vortices are not shed into the flow and the
separation region does not extend downstream. With further increase in the Reynolds number
the twin vortices elongate and a harmonic oscillation is observed in the wake region. At
Re 90 eddies start to shed alternatively from the upper and lower part of the cylinder and
form the so-called Karman Vortex Street. At this range of Reynolds number, the flow is
still laminar in the wake. As the Reynolds number increases (200 < Re < 300) the flow in
the wake becomes turbulent, while the boundary layer flow near the cylinder is still laminar
and remains laminar up to Re 3 105 (subcritical regime). We then have the critical
regime for 3 105 < Re < 3.5 105 where transition to turbulence spreads to the boundary

2 Laminar Vortex Shedding

layer region and the boundary layer becomes turbulent at the separation point at one side
of the cylinder while the boundary layer separation at the other side is still laminar. The
side where separation is turbulent occasionally switches from one side to the other. Sumer
[1] reports a non-zero mean lift on the cylinder because of this asymmetry. In supercritical
regime (3.5 105 < Re < 1.5 106 ) the boundary layer is turbulent at the separation point
for both sides but before the separation point it is partially laminar and partially turbulent.
Transition from laminar to turbulent happens at some point between the stagnation point
and the separation point. Then in the upper transition regime (1.5 106 < Re < 4 106 )
the boundary layer is completely turbulent for one side while on the other side it is partially
laminar and partially turbulent. Finally for Re < 4 106 boundary layer is fully turbulent for
both sides of the cylinder (transcritical). The above described stages of flow around cylinder
are depicted in figure 2.1.
Change of drag coefficient with Reynolds number can be explained using the above discussion
on flow regimes at different Reynolds numbers. Figure 2.2 shows drag coefficient as a function
of Reynolds number for a smooth cylinder. For Re > 3 105 transition to turbulence happens
and the separation point moves backwards which decreases the size of the wake, as a result the
total pressure force acting on the cylinder (and the drag coefficient) decreases abruptly.
The frequency of vortex shedding is described by the dimensionless Strouhal number, after
the Czech physicist Vincent Strouhal. Strouhal number is defined as:
fL
V

(2.2)

DR
AF
T

St =

where L is the characteristic length. For flow around a cylinder the diameter is taken as
the characteristic length. For 200 < Re < 3 105 Strouhal numbers remains approximately
constant and St 0.2 is a good estimate for it. Then there is a jump in Strouhal number at
transition Reynolds number. Figure 2.3 shows how Strouhal number changes with Reynolds
number for a smooth cylinder.

2.2 Numerics
Using discretization schemes the Navier-Stokes equations are transformed form partial
differential equations to algebraic equations. Different schemes are available to discretize
convection, diffusion and gradient terms. They have different properties in terms of accuracy
and stability and should be choose properly with respect to the specific problem. First order
schemes are in general too diffusive and would under predict forces and gradients.

Laplacian schemes
Laplacian terms are computed using Gaussian integration over the cell. In order to perform the
Gaussian integration values of the gradients at Gauss points should be calculated first. These
values are calculated via interpolation schemes and using the data at cell and face centers.
An interpolation scheme should also be assigned to the chosen laplacian scheme, usually a
linear interpolation is chosen. Other available interpolation schemes include: cubicCorrection,
midPoint, upwind, linearUpwind, limitedLinear, vanLeer, etc.

DR
AF
T

2.2 Numerics

Figure 2.1: Flow pattern at different Reynolds numbers, from [1]

2 Laminar Vortex Shedding

Figure 2.2: Drag coefficient as a function of Reynolds number for a smooth cylinder, from [2]

DR
AF
T

Gradient schemes
Gradient terms can be calculated by either Gauss integration or using leastSquares method.
Second and fourth order least square methods are available. In contrast to the Gauss
integration method, least square methods do not need an interpolation scheme. The schemes
available for discretizing gradient terms are listed in table 2.1. Gradient limiters can also be
used in discretizing gradient terms, they will avoid under shoots and over shoots in calculation
of the gradients. faceLimited schemes are in general more diffusive than cellLimited ones.

Divergence schemes
Similar to Laplacian schemes, Gauss integration is the only discretization method available
for discretizing convective terms. A proper interpolation method should be chosen. The
simplest method for calculating the divergence is the central difference scheme, which is
equivalent to Gauss integration with linear interpolator. This method is however not stable,

Table 2.1: Discretization schemes available for gradients

Discretization scheme
Description
Gauss <interpolationScheme>
Second order, Gaussian integration
leastSquares
Second order, least squares
fourth
Fourth order, least squares
cellLimited <gradScheme>
Cell limited version of one of the above schemes
faceLimited <gradScheme>
Face limited version of one of the above schemes

2.3 OpenFOAM

Figure 2.3: Strouhal number for a smooth cylinder, from 1

DR
AF
T

thus upwind scheme is normally used as the standard method for interpolation. Table 2.2
lists available interpolation schemes in OpenFOAM.

2.3 OpenFOAM
In the following the steps towards simulation of laminar flow around a cylinder are explained.
The schematic of the problem is presented in figure 2.4.

2.3.1 Preprocessing
Preprocessing mainly consists of preparing the blockMeshDict to generate the block-structured
mesh, setting the required boundary conditions in the 0 folder and setting the parameters in

Table 2.2: OpenFOAM interpolation schemes

linear
Second order, unbounded
skewLinear
Second order, (more) unbounded, skewness correction
cubicCorrected
Fourth order, unbounded
upwind
First order, bounded
linearUpwind
First/second order, bounded
QUICK
First/second order, bounded
TVD schemes
First/second order, bounded
SFCD
Second order, bounded
NVD schemes
First/second order, bounded

2 Laminar Vortex Shedding

Figure 2.4: Schematic presentation of the computational domain

.
contolDict. Based on what you want to simulate you might also need to modify the schemes
used within the solver for discretizing and solving the governing PDEs which is done in
fvSchemes file in system directory.
Meshing

DR
AF
T

The mesh can be generated using different blocking strategies. Here we use four blocks to
generate an o-grid type mesh around the cylinder. A fifth block is attached to the o-grid
mesh to extend the domain in the downstream region of the flow, figure 2.5. The four edges

Figure 2.5: Dividing the computational domain into 5 blocks

.
3 4, 4 6, 6 5 and 5 3 are straight lines by default. In order to represent cylinder
edges using them, they should be changed to arcs. It can be done in the edges section of the
blockMeshDict using arc command:
1
2
3
4

edges
(
arc 3 5 ( -1 0 0)
);

2.3 OpenFOAM
The first two numbers indicate the start and end point of the arc. Infinite number of circular
arcs can be defined between two points, we need a third point to specify the exact arc we
need. The coordinate of this third point follows after start and end point numbers in the arc
command, (1, 0, 0) in this case. There are other options for defining curved edges between
to vertices. Using splines for example, one can represent an airfoil profile using just 3 vertices
and defining 3 splines for edges connecting them. Available options for defining curved
edges in blockMeshDict are: arc, simpleSpline, PolyLine and polySpline. Their description is
available in OpenFOAM user guide.
Boundary conditions

DR
AF
T

Six patches are defined in this problem: inlet, outlet, top, bottom, cylinder, sides. Top and
bottom patches are defined to be far enough from the cylinder, they are defined as slip walls.
Since it is a 2D problem, side walls are defined in a single patch with type empty. The
boundary conditions used are summarized in table 2.3. The zeroGradient boundary condition
implies the gradient to be zero normal to the patch and slip is the same as zeroGradient
for scalar parameters while for vectors it implies fixedValue zero to the normal to the patch
component of the vector and zeroGradient to tangential components of the vector. Other
used boundary conditions are more or less self-explanatory. The magnitude of the velocity at
the inlet is set to a fixed value of 1m/s, to fulfill Re = 100 we can later play with dynamic
viscosity.
Table 2.3: Boundary conditions

BC
U

Patch
Boundary condition
inlet
fixedValue, uniform (1 0 0)
outlet
zeroGradient
cylinder fixedValue, uniform (0 0 0)
top
slip
bottom
slip
sides
empty
inlet
zeroGradient
outlet
fixedValue, uniform 0
cylinder
zeroGradient
top
zeroGradient
bottom
zeroGradient
sides
empty

controlDict
We have already explained the controlDict in our tutorial #3. Here we explain how to define
functions to calculate forces and force coefficients on a certain patch during run time. We
can later use these outputs to validate our simulation.
A function block can be defined at the end of controlDict. Functions perform a postprocessing step at the end of each time step. Different functions are available including:
faceZonesIntegration, fieldAverage, fieldMinMax, probes, etc.

2 Laminar Vortex Shedding

A generic function block looks like the following:
Listing 2.1: Definition of a general function in controlDict
functions
{
// a given name to distinguish between defined functions
name
{
// name of the function e . g . forces , probes ,...
type
< functionName >;
// name of function s library
functionObjectLibs ( " < libName > " ) ;
// control parameters for function output
outputControl
timeStep ;
outputInterval
10;
... // function - specific entries
}
}

In this tutorial we are going to use the two following ones:

forces: calculates the forces and moments acting on a list of patches by integrating
the pressure and shear forces acting on them. Additional entities like fluid density and
center of rotation of the body should be specified by the user. Defining a force function
in the controlDict is as follows:

DR
AF
T

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16

Listing 2.2: Definition of force function in controlDict

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17

forces
// given name
{
type
forces ;
functionObjectLibs ( " libforces . so " ) ;
outputControl
timeStep ;
outputInterval
10;
patches
pName
UName
rhoName
log

( " cylinder " ) ;

p;
U;
rhoInf ;
true ;
// true writes the data to the
// shell , false doesn t .
(0 0 0) ; // center of rotaion of the body
1.225;
// fluid density

CofR
rhoInf
}

forceCoeffs: calculates lift, drag and pitching moment coefficients for a patch list.
These coefficients are defined as:
Cl =

10

L
1
2
AU
2

(2.3)

2.3 OpenFOAM

Cd =

CM =

(2.4)

1
2
AU
2

(2.5)

1
2 c
AU
2

To calculate the coefficients the reference area and reference length (for moment
coefficient) should be specified by the user. By definition, the lift force acts perpendicular
to the velocity vector and drag acts parallel to it. These directions as well as pitching
moment direction are declared in function block:
Listing 2.3: Definition of forceCoeff function in controlDict
forceCoeffs
{
type
functionObjectLibs
outputControl
outputInterval

forceCoeffs ;
( " libforces . so " ) ;
timeStep ;
10;

patches
pName
UName
rhoName
log

( " cylinder " ) ;

p;
U;
rhoInf ;
true ;

liftDir
dragDir
CofR
pitchAxis

(0
(1
(0
(0

magUInf
rhoInf
lRef
Aref

1;
1.225;
1;
2;

DR
AF
T

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23

1
0
0
0

0) ;
0) ;
0) ;
1) ;

2.3.2 Solver
pisoFoam is a transient solver for incompressible flow. Both laminar and turbulent flows can
be simulated using pisoFoam. It is based on the PISO algorithm, which stands for Pressure
Implicit with Splitting of Operators, proposed by Issa in 1995. It is a pressure-velocity
coupling algorithm with a predictor step and two corrector steps. In the predictor step the
momentum equations are solved for an intermediate pressure field. The predicted velocity
field at this stage do not satisfy continuity condition. Then during the two predictor steps,
velocity and pressure fields are corrected in a way to fulfill both, momentum and continuity
equations. The PISO algorithm consists of the following steps:

11

2 Laminar Vortex Shedding

Prediction of the intermediate velocity field from momentum equations
PISO loop
Calculate mass fluxes on cell faces
Solve pressure field (nNonOrtcorr times)
Correct Velocity field
Repeat the PISO loop nCorr times
Solve the turbulence model
Proceed to the next time step
Schematic presentation of the algorithm can be seen in figure 2.6. The way the algorithm is
implemented in OpenFOAM can be seen in the source files of pisoFOAM solver:
Listing 2.4: Implementation of PISO loop in pisoFoam
// Pressure - velocity PISO corrector
{
// Momentum predictor
fvVectorMatrix UEqn
(
fvm :: ddt ( U )
+ fvm :: div ( phi , U )
+ turbulence - > divDevReff ( U )
);
UEqn . relax () ;

DR
AF
T

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40

if ( momentumPredictor )
{
solve ( UEqn == - fvc :: grad ( p ) ) ;
}
// --- PISO loop

for ( int corr =0; corr < nCorr ; corr ++)

{
volScalarField rAU (1.0/ UEqn . A () ) ;
U = rAU * UEqn . H () ;
phi = ( fvc :: interpolate ( U ) & mesh . Sf () )
+ fvc :: ddtPhiCorr ( rAU , U , phi ) ;
adjustPhi ( phi , U , p ) ;
// Non - orthogonal pressure corrector loop
for ( int nonOrth =0; nonOrth <= nNonOrthCorr ; nonOrth ++)
{
// Pressure corrector
fvScalarMatrix pEqn
(
fvm :: laplacian ( rAU , p ) == fvc :: div ( phi )
);

12

2.3 OpenFOAM

pEqn . setReference ( pRefCell , pRefValue ) ;

if
(
corr == nCorr -1
&& nonOrth == nNonOrthCorr
)
{
pEqn . solve ( mesh . solver ( " pFinal " ) ) ;
}
else
{
pEqn . solve () ;
}
if ( nonOrth == nNonOrthCorr )
{
phi -= pEqn . flux () ;
}
}
# include " continuityErrs . H "
U -= rAU * fvc :: grad ( p ) ;
U . c o r r e c t B o u n d a r y C o n d i ti o n s () ;
}
}
turbulence - > correct () ;

DR
AF
T

41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68

13

2 Laminar Vortex Shedding

Momentum predictor,
initial guess of U
First estimation of
U, interpolation of
fluxes on the faces
Solve the pressure
equation, correct
the flux on faces

Increase time

Yes

Correct non-orthogonality
No
Correct the velocity

Yes

More corrections required?

DR
AF
T

No

No

Solve turbulence model

End time reached?

Finalize the run

Figure 2.6: Piso algorithm

2.3.3 Post processing

In this part the results of the simulation are evaluated and compared with reference results
reported in the literature. The initial flow needs some time to develop around the cylinder,
what happens after a while is the onset of vortex shedding and development of the von
Karman vortex street (figure 2.7). The Periodic shedding of the vortices into the wake results
in an oscillating lift and drag force on the cylinder. The change in lift and drag coefficients in
time can be seen in figures 2.8 and 2.9. With the emergence of vortex shedding and as flow
develops itself around a cylinder the drag coefficient begins to oscillate about a mean value of
1.386 and the amplitude of lift coefficient is 0.358. They are in a very good agreement with
the reported values for drag coefficient and the amplitude of lift coefficient from [3] which are
1.39 and 0.35.

14

2.3 OpenFOAM

Figure 2.7: Velocity distribution around the cylinder and the von Karman vortex street

Cl

0.4
0.3
0.2
0.1
0
0.1
0.2
0.3
0.4
0

DR
AF
T

50

100 150 200 250 300 350 400

Time [sec]
Figure 2.8: Lift coefficient

The next quantity to be validated is the frequency of vortex shedding. It can be done by
Fourier analysis of the forces acting on the cylinder. Using Fourier transform the amplitude of
different frequencies present in a signal (here lift coefficient) are calculated (figure 2.10). As
can be seen in the figure the oscillations in lift coefficient is dominated by a single frequency
at f = 0.084 Hz, which is the frequency at which vortices are being shed into the wake. The
corresponding Strouhal number can be calculated from equation 2.2:
St =

fL
0.084 2
=
= 0.168
V
1

(2.6)

It matches very well with the reported value of 0.17 from [3].

15

Cd

2 Laminar Vortex Shedding

1.5
1.45
1.4
1.35
1.3
1.25
1.2
1.15
1.1
0

50

100 150 200 250 300 350 400

Time [sec]

0.14
0.12
0.1
0.08
0.06
0.04
0.02
0
0.02
0

DR
AF
T

Amplitude

Figure 2.9: Drag coefficient

0.05

0.1

0.15 0.2
0.25
Frequency [Hz]

0.3

Figure 2.10: Fourier analysis of lift coefficient

2.4 Exercise
Laminar pipe flow is another benchmark for laminar fluid flow. Use the same fluid setup, to
simulate laminar flow in pipes. Observe the development of velocity profile from the entrance
to the region of fully developed flow. The velocity profile, together with the entrance length
and pressure drop in the pipe could be good candidates to evaluate your results.

2.5 References
1. Sumer, B. Mutlu, et al. Hydrodynamics Around Cylindrical Structures. Singapore:
World Scientific, 2006
2. Fox, Robert W., et al. Introduction to Fluid Mechanics. Hoboken, NJ: Wiley, 2010

16

2.5 References

DR
AF
T

3. Durbin, Paul A., et al. Fluid Dynamics with a Computational Perspective. Cambridge:
Cambridge Univ. Press, 2007

17