2011 International Conference on Alternative Energy in Developing Countries and Emerging Economies
- 315 -
Abstract
-- For owners and manufacturers of wind turbines,
it is crucial that their machines operate inside guaranteed
power curves. In order to measure wind speed, standard
performance tests (IEC61400-12-1) use meteorological
masts. Draft IEC61400-12-2 recommends the use of nacelle
anemometry instead of the meteorological mast. While this
alternative simplifies the test process and reduces costs,
multiple factors influence the measurements taken at the
nacelle anemometer. The effects of rotor and nacelle are
among the most significant of these. Terrain is also a
significant source of uncertainty for the nacelle anemometry
technique and is discussed in this paper. Numerical 3D-
RANS calculations are carried out on two types of wind
turbines (Jeumont J48 and Nordex N80), each in an
atmospheric boundary layer. Flat terrain and two
escarpments are considered. Three types of effect on nacelle
anemometry are investigated: terrain slope, roughness and
hub height.
Index Terms
—
Escarpment, 3D numerical simulation,
nacelle anemometry, roughness, wind turbine
I.
I
NTRODUCTION
The wind turbine performance test is a valuable step
towards identifying problems linked to low energy
production. This test requires a high quality of measured
free stream wind speed. Although International Standard
IEC61400-12-1 describes such a technique, the use of a
meteorological mast complicates the task and increases
costs. To avoid these drawbacks, a nacelle anemometer
can be used to construct the power curve [1]. However,
flow at the position of the anemometer is affected mainly
by the shape of both the nacelle and the blade roots. Thus,
in order to be able to make use of nacelle anemometry,
one needs to know the relationship (NTF, Nacelle
Transfer Function) linking free stream wind speed
(FSWS) and nacelle wind speed (NWS). The terrain's
topography is also a significant source of uncertainty [2].
Flow over hills produces significant speed-up where its
maximum is reached at the crest [3]. RANS methods
predict this phenomenon fairly well on the lee side of a
hill [4], but not in the separated region. The inclined
airflow associated with sloped terrain can have a
significant impact on estimated annual production [5].
Surface roughness affects mainly shear and level of
turbulence intensity. In the case of a rough hill, speed-up
on the crest is higher in cases where the terrain is smooth
This work was supported by NSERC
’s
Wind Energy Strategic
Network and by the Canada Research Chair Program.
[6]. Furthermore, the point of reattachment also depends
on the type of terrain [7].
This paper aims to analyze the effects of sloped terrain on
nacelle anemometer readings. This was done via a
numerical evaluation of NTF for a Nordex turbine which
was installed on flat terrain as well as on two escarpments
(11% and 20%, respectively). For the flat terrain, free
stream wind speed was simply the speed prescribed at the
domain inlet. For the sloped terrain, a numerical site
calibration had to be accomplished before constructing
the NTF curve. The numerical calibration consisted in
modeling and simulating the escarpments that had no
wind turbine; this method was used to investigate the
ways in which that type of topography influences flow.
The second part of this paper analyzes the effect of
terrain roughness on flow in the vicinity of the nacelle,
particularly in the area where the anemometer is placed.
For this investigation, the turbine as well as the
escarpments from this study's Part One were used. Lastly,
the effect of hub height on NTF was evaluated by
positioning the Jeumont wind turbine at various levels on
a flat terrain.
II. M
ATHEMATICAL
M
ODEL
A. Governing equations
Numerical simulations were carried out by solving the
3D-RANS equations. Atmospheric flow was assumed to
be steady, incompressible, turbulent and without thermal
stratification. To close the system of equations, the
k-
ε
model was used. Despite its isotropy, the
k-
ε
model has in
fact been used extensively in the field of wind energy
with various sets of values for the empirical constants.
For this study, the set used by Crespo [8], calibrated for
neutral atmospheric boundary layer, was chosen.
B. Boundary conditions
Boundary conditions were imposed on a cube-shaped
domain (Fig. 1). At the inlet of the domain, velocity and
turbulence profiles were imposed. The logarithmic profile
was chosen, taking the hub height of the turbine as
reference height. At the outlet, the normal gradient of all
variables was zero. At the lateral surfaces, an absence of
transversal flow was assumed, with normal zero gradients
for all variables. On the ground and in the topmost areas,
shear stress was imposed. A logarithmic profile on a local
basis was assumed in the cells near these boundaries. The
centroid of the second cell closest to the boundary was
used to evaluate friction velocity.
K. Ameur and C. Masson
École de Technologie Supérieure, Département de Génie Mécanique, (
Canada
)
Wind Turbine Testing Using Nacelle
Anemometry: Ground Effects
