2011 International Conference on Alternative Energy in Developing Countries and Emerging Economies
- 316 -
Fig. 1. Computational domain with boundary conditions.
C. Wind turbines
The first turbine used in the calculations was a 2.5MW
Nordex N80, with a diameter (
D
) of 80m and a hub
height (
H
hub
) of 80m. The rotor was modeled using the
concept of the actuator disk [9], which consists of a
porous surface where the effects of the blades are applied
as source terms in the momentum equations. The effect of
the lifting portion of the blade was estimated using the
thrust coefficient provided by the manufacturer. For the
blade root, considered as a circular cylinder, the induced
axial force was evaluated using the BEM (Blade Element
Momentum) method. The second turbine was a 750kW
Jeumont J48, with a diameter of 48m and a hub height of
46m. Since the geometric and aerodynamic
characteristics of the blade were known, the BEM method
was applied over the entire rotor.
For both turbines, nacelle geometry was fully represented
in the mesh (Fig. 2). Unlike the case for the Nordex, the
geometry of the Jeumont nacelle was highly symmetrical
surrounding the axis of rotation of the rotor. The wall of
the nacelle was considered smooth, where shear was
evaluated via the standard Launder and Spalding law of
the wall [10]. The tower was not represented.
III. N
UMERICAL
M
ODEL
Fluent 6.3 was used to solve the RANS equations
governing flow. The treatment of pressure-velocity
coupling was carried out using the SIMPLE algorithm.
The convective terms were discretized by the third-order
quick scheme. For diffusion terms, a second-order
centered scheme was used.
IV. C
OMPUTATIONAL DOMAIN AND MESHES
The computational domain took the shape of a
rectangular parallelepiped. A grid independence study
was conducted on the 20% escarpment using an
operational turbine. The dimensions of the resulting
domain were: 30
D
, 6.25
D
and 20
D
respectively, towards
X, Y and Z. For each escarpment studied, two meshes
were created, one with a wind turbine and the other with
no turbine.
The meshes were structured and consisted essentially
of hexahedral cells. A refinement of the mesh was
operated near the walls of the nacelle, at ground level and
around the beginning and end of the escarpment.
Additionally, the mesh was stretched horizontally and
vertically towards all boundary surfaces. To save on the
number of cells used, only half of the wind turbine was
taken into consideration (Fig. 2), since there was no yaw.
This yielded an average of 2.10
6
cells per grid. This
number of cells was obtained following a grid
convergence study performed on three meshes with
increasing refinement.
Fig. 2. Vicinity of the nacelle and rotor for (a) Nordex N80 and (b)
Jeumont J48.
The various slopes of the ground were obtained by
varying the horizontal length of the escarpment (Fig. 3a)
while maintaining a fixed height (
H
esc
) of 0.625
D
. The
Nordex turbine was placed at the middle of the
escarpment with a constant hub height. The roughness
height of the ground in the first part of this study was set
to 0.05m. This value was the same as that seen in
experimental results obtained from ECN (Energy
research Centre of the Netherlands). The results relating
to greater distances were rendered dimensionless by
H
hub
,
results related to the vicinity of the nacelle by
L
nacelle
and
H
anemo
(Fig. 3b).
Fig. 3. Description of the (a) escarpment and (b) the nacelle vicinity.
V. R
ESULTS AND
D
ISCUSSION
A. Slope terrain effects
Terrain with no wind turbine
Figs. 4 and 5 provide information on the inflow
conditions for mean flow and turbulence, respectively.
Fig. 4 shows the effect of the escarpment on the axial
(a)
(b)
