Challenges in modelling offshore wind

1
Challenges in modelling offshore wind how to
address them using observation

• Idar Barstad
• idar.barstad_at_uni.no

2

• Improvements of resource estimates and forecast
  of energy yield rely primarily on the quality of
  numerical models and their input data.
• There are many ways to set up the model suite,
  and the numerical model tool is normally tailored
  for generic use.
• The system can sometimes be inherently
  unpredictable

3
Future wind power potential in Europe
- Arpege/Ifs T159L60c3 -- (1972-2001)
N/F --(2020-2049) R1-4 (A1B) SST ERA40
delta from CGCMs (variability from
ERA40 gtget more realistic sea ice)
Barstad et al. (2012), J. Renewable Energy
Barstad et al. (2008), Clim.Dyn.
4
Current wind climate
5
Future power potential(2020-2049)
Fractional power potential in reference to
(1972-2001)
Black line1.0
6
Example 12UTC 29Feb 2008
The effect of surface waves
7
WRF roughness length (m) after 12 h
Work by Alastair Jenkins, Alok Gupta, John
Michalakes (NREL), Idar Barstad
8
The effect on U10 (wind speed) in WRF
2-way coupling Significant impact on the wind
field!
After 12 hrs simulation
9
High resolution downscaling (9-3km)
10
High resolution downscaling (9-3km)
11
BIAS
(Courtesy M Zagar, Vestas
12

• Valentia Irland
• - 6700 cases over 10 yrs (2000-2010)
• - Conditions for lee waves
• 15-20 of the time

13
Wind turbine drag in WRF from V3.3

• Drag from turbines in a single cell, distributed
  over several layers
• - gt works on both the TKE and the momentum eqs.

(Blahak et al. 2010)
14
Simulation of a wind farm100 x 5MW wind
turbines
Fitch et al. (2012) MWR

• Q
• How sensitive is the power output to the
  atmospheric characteristics?

15
The principle
Generation of pressure gradients by wind farm -
? increases with height under typical stable
conditions - As air lifted over farm, lower ? air
brought up from below - This creates cold anomaly
aloft and thus high pressure anomaly below
gt pressure gradients deflect wind.
Slide 3
16
Work by Fitch and Barstad
Slide 11
17
Demonstration of the WRF-turbine dragat Dogger
bank
(9km-3km-1km)
15 x15 km with a 5 MW turbine in each grid cell
18
Dogger bank wind farm
(15 x15 km with a 5 MW turbine in each grid cell)
1km domain wind and wind speed
1km domain wind, wind speed vorticity
Conclusions - 10 effect on wind speed (up to
60 on power) - Long wakes
19
Reduced model single farm
Same model as in Smith (2009)
20
Reduced model single farm
The effect of the inversion strength
double inversion strength
21
Two farms reduced model
Second farm (Lx10km)
Present in all runs (Lx10km)
Utop15 m/s UBL8.5 m/s H500m dth/th0.01
22
Two farms reduced model
Second farm (Lx10km)
Present in all runs (Lx10km)
23
Conclusions

• Models may produce data, but we have to be
  critical to their results
• Models may be tailored to your specific needs -gt
  talk to an expert!
• Observational campaigns should be design to
  address scientific questions. Do we have these
  questions?

24
Thank you for your attention!Idar.Barstad_at_uni.no
Alaska
NOAA / MODIS 23 MAR 2010, Aleutians Islands