Saarblitz, Wind farm in HDR, flickr.com, creative commons by-nc-sa 2.0

Lillgrund 360 Efficiency

Managed by

Scope

The benchmark is open to participants of the Wakebench project using wake and, possibly, atmospheric boundary layer models. This is a case based on the actual operational Lillgrund wind farm in which there are multiple turbines interacting within an array.  This benchmark aims to test a wake/atmospheric model to reproduce the wind plant efficiency observed at Lillgrund over the full wind rose.

Objectives

Demonstrate how wake models perform and capture the wake formation and merging process in the presence of atmospheric shear and turbulence within a large modern wind farm composed of modern multimegawatt turbines. 

Data Accessibility

Brief description about the accessibility of the data

Input data

The conditions for simulating the Lillgrund_360_Efficiency case are:

  • Location: 55.52° N, 12.78° E
  • Wind directions: Full 360° wind rose starting at 0° in 3° increments and ± 1.5° bins.
  • Hub height turbulence intensity:  This quantity is defined as ‹u’u’›1/2/U0 where ‹u’u’›1/2 is the streamwise velocity variance and U0 is the mean hub-height wind speed.  It varies over the wind rose, so see Fig. 1 and Table 1 below for a description. If you are unable to use a direction-varying turbulence intensity, please use a uniform turbulence intensity of 6%.
  • Surface roughness:  z0 = 0.0001 m unless a different value is needed by your simulation method to produce the desired hub-height turbulence intensity.
  • Hub height mean wind speed: 9.0 m/s.
  • Stability: Assume neutral stability because stability is not given by Dahlberg (2009) or Bergström (2009) and data is averaged over a long time period.
  • Turbine model:  Siemens SWT-2.3.93.
    • The basic turbine characteristics of coefficient of power and thrust versus wind speed are given in the test case guide (Figure 2) and in the file SWT-2.3-93.txt. 
    • If a more detailed turbine model is used that requires blade and control system information, a detailed generic turbine characterization representative of the actual turbine and tuned to the Cp and CT data given in SWT-2.3-93.txt is given by Churchfield (2013).

Figure 1   Average turbulence intensity versus wind direction (from Bergström, 2009).

Table 1 Average hub height (65 m) turbulence intensity versus wind direction.

Validation data

The validation data consists of farm efficiency versus wind direction as shown in Figure 2 below

Figure 2   Wind farm efficiency versus wind direction (from Dahlberg, 2009).

Model runs

The participant should use a domain size sufficient for his or her model to work properly and devoid of spurious boundary effects.  Please run the simulations in 3° bin widths if possible.  For more computationally expensive methods, choose at least one low efficiency (one could use the 120º or the 300º case from the previous set of benchmarks) and one high efficiency wind direction (for example, the 315º case).  To properly capture efficiency, the entire wind plant should be modeled.

Output data

Output data to be reported are wind plant efficiency based on time-averaged power production data from each turbine where the averaging period is at least 10 minutes.  The normalization used in the efficiency calculation is direction dependent.  Use the following formula to calculate the efficiency

where θ is the wind direction, P is power and an overbar denotes the time-average, the subscript i denotes turbine number and is a set containing all 48 turbines, and the subscript j denotes turbine number and is a set containing only the turbines used for normalization, and Nj is the number of turbines in the set j.  The set j is dependent on wind direction, as shown in Fig. 3.  For example if the wind direction is in the range 351°< θ < 81°, then the set j contains turbines 1, 8, 16, 24, and 31, which are the most upwind turbines for that wind direction range.  Table 2 defines the set j of turbines for four different wind direction ranges covering the entire wind rose.

Table 2 Turbines used for normalization in efficiency calculation.

Figure 3   A diagram of turbines used for normalization in efficiency calculation depending on wind direction.

Please report your results using the following file format.

  • Efficiency data

Lillgrund360Efficiency_UserID_ModelID_run#_efficiency.txt

A multiple column list of efficiency based on time-averaged power and the formula given in Equation 1 with normalization from Table 2.  Each row (below the first row header) corresponds to a different wind direction.  An example is shown below.

windDir(°), efficiency
0, 0.756

299, 0.563
300, 0.548
301, 0.564

359, 0.754

References

Bergström, H., “Meteorological Conditions at Lillgrund,” [online report] Vattenfall Vindkraft AB, 6_2 LG Pilot Report, Mar. 2009, URL: http://www.vattenfall.se/sv/file/16_Meteorological_conditions.pdf_ 16614584.pdf  [cited 27 June 2012].

Churchfield, M. J., “Generic Siemens SWT-2.3-93 Specifications,” informal NREL document, 2013.

Dahlberg, J.-Å., “Assessment of the Lillgrund Wind Farm: Power Performance Wake Effects,” [online report] Vattenfall Vindkraft AB, 6_1 LG Pilot Report, Sept. 2009, URL:http://www.vattenfall.se/sv/file/15_Assessment_of_the_Lillgrund_ W.pdf_16596737.pdf  [cited 27 June 2012].

Jeppsson, J., Larsen, P.E., Larsson, Å.  “Technical Description of Lillgrund Wind Power Plant,” [online report] Vattenfall Vindkraft AB, 2_1 LG Pilot Report, Sept. 2008, URL:http://www.vattenfall.se/sv/file/2_Technical_Description_Lillgru_8459792.pdf_16611908.pdf  [cited 27 June 2012].