”K-epsilon (k-) Turbulence Model”

Gabriel F.R Lopes1

[email protected]
Ph.D. Nelson Moraga Benavides
January 21-2019

Abstract— In this project we will be able to use the K-epsilon

(k-) turbulence model,to simulate mean flow characteristics for
turbulent flow conditions. This model focuses on the mecha-
nisms that affect the turbulent kinetic energy. A problem related
to this model will be presented using FORTRAN(SAINTS)-
TECPLOT360-FLUENT.

I. I NTRODUCTION
F igure1
K-epsilon (k-) turbulence model is the most common
model used in Computational Fluid Dynamics (CFD) to sim-
ulate mean flow characteristics for turbulent flow conditions. To validate the results obtained by the Saints program,
It is a two equation model that gives a general description of we will perform the same exercise with the identical
turbulence by means of two transport equations(PDEs).The measurements in the Fluent program, using a new, more
original impetus for the K-epsilon model was to improve refined 100x100 mesh, these models will be represented
themixing-length model, as well as to find an alternative to below:
algebraically prescribing turbulent length scales in moderate
to high complexity flows. The first transported variable is
the turbulence kinetic energy (k). The second transported
variable is the rate of dissipation of turbulence energy
(). This model focuses on the mechanisms that affect the
turbulent kinetic energy. The mixing length model lacks this
kind of generality. The underlying assumption of this model
is that the turbulent viscosity is isotropic, in other words, the
ratio between Reynolds stress and mean rate of deformations
is the same in all directions.Next, a problem that is found in
Nakayama’s book using the k-epsilon turbulence model will
be solved

II. P HYSICAL L AYOUT

A. Applications of Program ”Saints” To Complex Turbulent Figure 2
Flows:
The physical situation considered here is illustrated
in Figure 1. A fluid of uniform temperature, velocity,
turbulence kinetic energy and its dissipation rate, enters into
a plane channel of 2 (m) height. The channel inlet is located
4 (m) upstream of the backward facing step of height. The
lower walls is assumed to be adiabatic. Note that the flow
is turbulent, the thermal properties of the fluid are:

ρ=1 [kg/m3 ];
v = 2.2 ∗ 10− 5[m2 /s];
Cp = 1000[J/Kg ∗ C];
P r = 0.71; Figure 3
Re = 4.6 ∗ 104
 Figure 6
Figure 4
In carrying out said procedure, it was defined that the
III. R ESULTS P RESENTATION AND DISCUSSIONS program should perform a total of 200 iterations, delivering
the values of K, Epsilon and their different velocities in a
In this section, results obtained for the problem will be graph, which is presented below:
presented.

A. Applications of Program ”Saints” To Complex Turbulent

Flows:
First the problem was made through the Saints program,
obtaining results for the speeds in the X direction. For
this case, a 32x32 mesh was used. In the results it can
be seen that the maximum speed is at the entrance of the
fluid and, as its trajectory passes, this speed decreases,
it can also be seen that the values of the velocities on
the walls are 0 (m / s) that complies with what has
been studied. The results are shown in the following figure.

Figure 7

To make another comparison and be sure that what is

represented in Fluent is correct, the values obtained for the
pressures in both programs are presented, it can be seen
that the maximum and minimum values of the pressures
are close to the results obtained, the difference observed is
that the length of the problem performed in Fluent is much
greater than in Nakayama’s book, so the results are identical
up to a certain value of L, the results are presented below:

Figure 5

In order to corroborate the results obtained with the

Saints program, the same problem was performed in Fluent
with the same dimensions and values, but with a more
refined and different 100x100 mesh, it can also be seen in
the figure that the results obtained are relatively identical to
the performed previously, the value of the maximum velocity
is also found at the entrance of the fluid and that when
the movement of the fluid elapses, its velocity decreases. Figure 8
 by excessive kinetic energy in parts of a fluid flow, which
overcomes the damping effect of the fluid’s viscosity. For
this reason turbulence is easier to create in low viscosity
fluids, but more difficult in highly viscous fluids. In general
terms, in turbulent flow, unsteady vortices appear of many
sizes which interact with each other, consequently drag due
to friction effects increases. The temperature values were
different from the book, since different values of initial
temperatures were used, but that the distribution presents
the same senses as Nakayama.
Figure 9
V. A PPENDIX
A final comparison was made with the temperatures,
the same initial values were not used as the book of Standard K- Turbulence Model:
Nakayama, but it can be concluded that both have the same
temperature distribution along their walls, the results are The exact k- equations contain many unknown and
presented below: unmeasurable terms. For a much more practical approach,
the standard k- turbulence model (Launder and Spalding,
1974) is used which is based on our best understanding
of the relevant processes, thus minimizing unknowns and
presenting a set of equations which can be applied to a
large number of turbulent applications

For Turbulent Kinetic Energy k:

Figure 12

For Dissipation:
Figure 10

Figure 13

Where µi: Represents velocity component in corresponding

direction
Eij : Represents component of rate of deformation
µt: Represents eddy viscosity.

Figure 11 Figure 14

The equations also consist of some adjustable constants.The

values of these constants have been arrived at by numerous
IV. C ONCLUSIONS iterations of data fitting for a wide range of turbulent flows.
These are as follows:
In this report it can be concluded that the values obtained
by the student using the Fluent program approach and
improve the one made in the Saints program, obtained
from Nakayama’s book. An improvement was made in the Figure 15
meshes of this problem, since when creating a mesh with
more divisions, values closer to the reality of the problem
are obtained. It can be concluded that turbulence is caused
Applications: there is an area close enough to the wall to be satisfied:

The k- model has been tailored specifically for planar

shear layers and recirculating flows.This model is the most
widely used and validatedturbulencemodel with applications
ranging from industrial to environmental flows, which Figure 18
explains its popularity. It is usually useful for free-shear
layer flows with relatively small pressure gradients as well But far enough away so that turbulent tensions dominate
as in confined flows where the Reynolds shear stresses over those due to viscosity, in which:
are most important.It can also be stated as the simplest
turbulence model for which only initial and/or boundary
conditions needs to be supplied. This model would be
an inappropriate choice for problems such as inlets and
compressors as accuracy has been shown experimentally Figure 19
to be reduced for flows containing large adverse pressure
gradients. The k- model also performs poorly in a variety of Where β is a constant, of value equal to 5, which
important cases such as unconfined flows,curved boundary has been determined experimentally.This area is known as a
layers, rotating flows and flows in non-circular ducts. logarithmic layer or inertial sublayer. Note that the functions
of the equations do not depend on the shape of the conduit,
Turbulent Flows In The Proximity Of a Wall (Law of so they describe a universal law of flow in the vicinity of
the wall): solid surfaces, called the law of the wall.Due to the behavior
of the turbulent flows in the vicinity of a wall that has just
At high Reynolds numbers, turbulent flows tend to be been described, the application of a non-sliding contour
macroscopically independent of viscosity (although not on condition on the walls and the integration of the closing
smaller scales). However, viscous effects affect movement in models of two equations through the viscous sublayer
the vicinity of a wall in a very important way. Consider the provides unsatisfactory results.
flow in a two-dimensional pipe with smooth walls (without
roughness). Far from the walls, in most of the conduit, the This is due to the inability of these models to predict
Reynolds tension is much greater than µ*du/dy, while in an acceptable value of the constant β in the law of
their proximity these two terms are made of the same order, the wall of the equation. For this reason it is generally
the first tending to be annulled in the wall due to that the used to impose boundary conditions on the walls for the
presence of this inhibits the turbulent fluctuations. It is easy speed, K and  by means of the so-called wall functions.
to demonstrate that in an area adjacent to the wall, called This procedure is based on using the law of the wall to
a viscous sublayer, where the shear stress can be assumed relate speed and shear stress. Once the friction velocity is
to be uniform and equal to the tension in the wall and the known, the values of K and  can be determined at the
turbulent tensions are negligible compared to those due to points of the calculation grid closest to the wall by means of:
viscosity, the velocity of the fluid increases. linearly with
the distance to the wall, according to the law:

Figure 16

Being u* the so-called friction velocity, which can be

interpreted as a typical value of the turbulent agitation
velocity, and τ p the shear stress in the wall. Experimentally Figure 20
it has been found that this linear velocity distribution is
maintained up to a dimensionless distance from the wall: One of the drawbacks of this method is that the numerical
solutions obtained are generally sensitive to the position of
the chosen point to apply the wall functions with respect
to its surface. An alternative to the use of wall functions
Figure 17 consists in the introduction of corrections due to the
viscosity in the model K-, allowing its integration through
which can be considered as the limit of the viscous the viscous sublayer.
sublayer. From an approximate value of Y+=30 up to an
upper limit of Y + that increases with the Reynolds number,
 R EFERENCES
[1] NAKAYAMA’S BOOK
[2] FORTRAN-SAINTS
[3] TECPLOT 360
[4] https://en.wikipedia.org/wiki/K-epsilonturbulencemodel
[5] https://www2.uned.es/ing-fluidos/IntroMF/node78.html