Professional Documents
Culture Documents
MARSEILLE
20
10
0
0 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
u [m/s]
Warning: Ieff exceeds IEC61400-1(ed.3) class b design limit
20
TI [%]
10
0
0 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
u [m/s]
Conclusion: Ieff within IEC61400-1(ed.3) class b design limit
Figure 2: Example of wind-speed dependent effective TI (dots) and IEC design references a) under normal opera-
tion and b) with turbine shut-down rules.
the upwind wake velocity deficit. There will, Weibull probability and linear power variation has an
however, be no velocity deficit contributions, see analytical solution and with cancelling terms it leads
equation 4, for shut-down turbines. to the sum
The two plots in figure 2 shows effective TI calcu- N
X −1
lated with and without turbine shut-down rules. The P = ∆i [P (ui+1 ) − P (ui )] (7)
strategy is very successful in this imaginary example, i=0
where turbines are deployed unusually close with dis- where the ∆i factor is
tances of about two rotor diameters. (
Gk (αi+1 )−Gk (αi )
αi+1 −αi 6 αi
for αi+1 =
∆i =
7 Regulated energy production − exp (−αi ) for αi+1 = αi
It is of interest to estimate the cost of wind sector man- using αi,j = ui /Aj for dimensionless speed and ex-
agement in terms of lost production. For this purpose pressing power-curves discontinuities as two records
we review the basic formula for accumulated energy with identical wind speed, Pi+1 6=Pi for αi+1 = αi .
production, used in WAsP [7] Gk (α) involves the incomplete gamma function.
Z ∞
Gk (α) = k −1 Γ k −1 , αk
P = P (u)p(u) du (5)
0 Z αk
−1
−1 −1 −t
The power curve P (u) is approximated by piecewise =k tk e dt (8)
0
linear variations
Production estimates with corrections from wake ef-
Pi+1 − Pi
P (u) = (u − ui ) + Pi fects and turbine shut-down rules could be written
ui+1 − ui
Z 2πZ ∞
for ui ≤ u < ui+1 (6)
P = δ(u − ∆u, θ)
0 0
and the wind-speed probability is expressed by sector-
wise Weibull distributions. The combination of · P (u − ∆u)p(u, θ) du dθ (9)
Here we introduce an indicator function δ(u, θ) which
is either unity or zero for an operating or shut-down
turbine. The integration variable is the ambient wind
speed u but both power curve and shut-down rules are
functions of wind speed corrected for wake velocity
deficit u − ∆u. The integration is not much different
from usual, as we can use an apparent power curve
P 0 (u, θ) = δ(u, θ)P (u).
In WAsP, the annual energy productions with and
without wake losses are called Gross AEP and Net
AEP, respectively. In WAT, the annual energy pro-
ductions with both wake losses and shut-down rules
is called Regulated AEP, and it will be lower than
the Net AEP for a regulated turbine but sometimes
slightly higher unregulated neighbour turbines, as Figure 3: Various sectors for IEC 61400-12-1 terrain
they are exposed less turbulence. assessment
8 Software nation of power data from the tested turbine and wind
data from a nearby reference mast. This is described
The above wake calculations are implemented in our in IEC 61400-12-1[8] and WAT supports some assess-
Windfarm Assessment Tool (WAT). WAT input is ments of that standard.
generated by a WAsP Engneering script, which cal-
culates and reports site-specific wind conditions. This • IEC 61400-12-1 Annex A includes rules for al-
script also calls WAsP for prediction of the mean wind lowable measurements sectors. Wind sectors are
climate so a twin WAsP/WEng license is required. excluded if measurements are disturbed at either
WAT is in itself free and can be downloaded from the tested turbine or the reference mast. Angles
www.wasp.dk/products/wat. of excluded sectors are prescribed by simple for-
mulae. Obstacle geometry is also needed, and
for this purpose WAT imports the WAsP obstacle
9 Terrain assessment file.
• IEC 61400-12-1 Annex B includes rules for al-
9.1 Complex-terrain factors lowable terrain complexity. If the terrain is too
IEC 61400-1 has an additional rule, which we have complex, a second met mast must be installed at
not discussed sofar. It declares that in complex terrain the exact turbine position prior to turbine instal-
the standard deviation of the longitudinal TI must be lation. After sufficient data collection, a statisti-
multiplied by a complex-terrain safety factor cal correlation between wind speeds at measure-
s ment mast and turbine position is established.
2 2
1 σ2 σ3 This correlation is later used to correct wind
Cct = 1+ + (10) measurements for the power performance mea-
1.375 σ1 σ1
surements. This method is called site calibra-
The terrain around a turbine site is categorization as tion. It implies planning at an early stage and ad-
complex according to certain rules considering maxi- ditional field work, but fortunately it is allowed
mum deviations between local terrain elevation and a to skip it if the terrain is less complicated than
plane fitted to the terrain around the turbine as well as specified in Annex B. The terrain assessment
the slope of that plane. WAT base the complex-terrain rules are similar to the complex-terrain indica-
assessments on a terrain model dumped from WEng tion rules in IEC 61400-1, except that for per-
and evaluates Cct by directional turbulence charac- formance measurement, the tested terrain is lim-
teristics in the WEng script. In WAT, it is optional ited to annular sectors instead of a circle. The
whether to apply Cct factors, as it could be argued terrain fitting method used in WAT is described
that this correction is unnecessary because the Mann in the Appendix of the present paper. If terrain
turbulence model adequately redistribute turbulence rules are exceeded by less than 50% it is allowed
energy among the three velocity components. to use modelled flow corrections estimated by a
flow model, e.g. LINCOM.
9.2 Site calibration
Figure 3 indicates various sectors as defined by
On-site measurements are used to document wind turbine layout and mast position according to IEC
farm power performance. This is done by a combi- 61400-12-1 Annex A and B. WAT fits planes to
these sectors and determine whether site calibration is Appendix
needed. The user is allowed to drag the mast around
in search for favourable measurement positions. IEC 61400-12-1 Annex B makes a terrain assessment
in which local terrain deviations in a annular sector is
tested from a fitted plane forced through the base the
10 Conclusions turbine tower z0 . The integral of the squares devia-
tions is
All wind conditions needed for an IEC 61400-1 site Z θ2Z r2
assessment can be found by post-processing of flow ∆= r [z (r, θ)
and turbulence model results followed by some ad- θ1 r1
ditional wake modelling. We work with WAsP and − (z0 + a · r sin θ + b · r cos θ)] dr dθ
2
(11)
WAsP Engineering results and implement additional
wake calculations in a new program called Windfarm where a and b are slopes in x and y directions, and the
Assessment Tool (WAT). Effects of turbine shut down sector is confined by the angles θ1 and θ2 and radii r1
for turbulence mitigation and some support for plan- and r2 . Minimizing ∆ with respect to slope gives the
ning of an IEC 61400-12-1 site calibration are in- solution
cluded. WAT is still under development, and a better
match of the neutral-stability TI of WAsP Engineering a −1 T +S −C Da
= 2 (12)
with the 90% percentile of all turbulence conditions, b T − C 2 − S 2 −C T − S Db
as required by IEC 61400-1 Ed. 3, is highly desirable.
using the short-hand notation
So is possible use of measured turbulence data.
T = 2θ2 − 2θ1 (NB: Angles in [rad] here)
S = sin 2θ2 − sin 2θ1
References C = cos 2θ2 − cos 2θ1
Z θ2Z r2
[1] IEC 61400-1 Ed. 2, wind turbine safety system - Da = [z (r, θ) − z0 ] 2r2 sin θdr dθ
Part 1: Safety requirements, 1999. θ 1 r1
Z θ2Z r2
[2] IEC 61400-1 Ed. 3, wind turbines - Part 1: Design Db = [z (r, θ) − z0 ] 2r2 cos θdr dθ
requirements, 2005. θ1 r1
[3] P. Astrup and S. E. Larsen. WAsP Engineering The integrals Da and Db are found numerically.
flow model for wind over land and sea. Technical Finally max |z0 + a · r sin θ + b · r cos θ − z (r, θ)|
Report Risø-R-1107(EN), Risø National Labora- within the annular sector is determined.
tory, 1999.