Back to the programme printer.gif Print



Tuesday, 11 March 2014
14:15 - 15:45 Aspects for offshore and complex terrain
Science & Research  


Room: Llevant
Session description

Siting in complex terrain is still a challenge for wind energy developers. This session will shed light on questions including:

What is the best way to estimate the energy resource over hilly and forested terrains? Are linearized models still useful and are computational fluid dynamics (CFD) models mature enough? What is the best way to use meso-scale models for wind energy resource estimation offshore? How does atmospheric stability affect turbulence?

Learning objectives:

  • Judge the performance of linearized flow models in comparison with CFD model for wind resource estimation in complex terrain
  • Understand how meso-scale models are best used for offshore annual energy production (AEP) estimation
  • Appreciate how atmospheric stability affects turbulence over forests and how to use standard measurements to estimate the stability
  • Get an insight into state-of-the-art meso-scale modelling for wind energy resource estimation
Lead Session Chair:
Jakob Mann, DTU Wind Energy

Co-chair(s):
Joachim Peinke, Uni Oldenburg, Germany
Simon Watson Loughborough University, United Kingdom
Co-authors:
Simon Watson (1) F P James Hughes (1)
(1) Loughborough University, Loughborough, United Kingdom

Printer friendly version: printer.gif Print

Abstract

Mesoscale modelling of the UK offshore wind resource

Introduction

Knowledge of the wind conditions at a potential offshore wind farm site is key in reducing investment risk. This is normally done through the use of large meteorological masts. However, the increasing scale of the turbines offshore requires higher and more expensive masts, driving interest in the use of alternatives to extend accurate assessment of the resource. This work examines the use of a mesoscale model for assessing the wind resource at UK offshore sites. A comparison is made with existing data and a projection is made of the wind conditions and variability at a potential UK Round 3 site.

Approach

This research involves the use of the Weather Research and Forecasting (WRF) mesoscale model [1, 2] to assess the wind conditions at selected sites in UK offshore waters. Specifically, the Advance Research WRF model core (ARW) is used in this work. The accuracy of the model is assessed in a number of ways: 1) Through application of several planetary boundary layer (PBL) schemes, both individually and as an ensemble; 2) through the use of time-step ensembles; 3) by the use of different timescale filters; 4) through the use of model ‘nudging’ using nearby observations. Each model run has its boundary conditions set using output from the National Centers for Climate Prediction (NCEP) Climate Forecast System Reanalysis (CFSR) [3]. This research is therefore concerned with how well a mesoscale model can downscale global forecast analysis data. A comparison is made between model output and observations from meteorological masts at Scroby Sands off the east coast of the UK and two masts at Shell Flats off the north-west coast. Model performance is assessed in terms of ability to predict: 1) mean wind speed; 2) wind direction; and 3) atmospheric stability. Recommendations are made in terms of how best to use the model for offshore wind resource prediction. Finally, a projection is made of the wind conditions at a future potential offshore wind farm site in the UK Round 3 Dogger Bank development zone. The variation in synoptic wind conditions across a large hypothetical 1.2GW wind farm in this area are also assessed including maximum expected wind speed and wind direction differences across the wind farm.

Main body of abstract

The WRF model was run using three levels of domain nesting with grid resolutions of 18km, 6km and 2km. Model runs were initialised using six-hourly data from the 0.5 degree x 0.5 degree CFSR product. Initial benchmarking was carried out against the Scroby Sands mast located ~3km from the coast. A comparison was then made between model output and observed data at the two offshore masts at Shell Flats (14km and 20km from the coast). Finally, a projection of wind conditions was made at a site in the UK Round 3 Dogger Bank development zone.

Table 1 shows a comparison between different studies using mesoscale models to predict offshore wind resource [4-9].


Table 1: Comparison between studies to predict offshore wind resource.

This study, based on the work of [10], shows a level of correlation with offshore sites which compares well with the other studies with an optimised RMSE of 2m/s and a best correlation of 0.9. It can be seen that the correlation with the observations at Shell Flats is better than that at Scroby Sands and this most likely reflects the relative importance and difficulty of modelling the land/sea effects which will be strongest for Scroby Sands.

Figure 1 shows a comparison between predicted wind speed at the Shell Flats Mast 2 and observations at 40m above sea level based on approximately 18 months of data.


Figure 1: Comparison between predicted and observed wind speed (in m/s) at Shell Flats Mast 2 at 40m above sea level.

There is relatively good agreement between the two wind roses, though there does seem to be an overall turning of the wind directions between the model and observations. The reasons for this are not clear, however, the data from Mast 1 ~5km from the Mast 2 shows a wind rose more consistent with that predicted. This suggests either a localised effect which has not been captured by the model or an offset in wind direction measurements at Mast 2.

The effect of nudging the model using wind speed measurements was investigated. For a two-month period, the WRF model was used to predict the conditions at Shell Flats Mast 2 using CFSR data only and then run again using additional wind speed data input from Mast 1. Without nudging the correlation coefficient was found to be 0.74. The additional of the nudging data from Mast 1 improves the correlation to 0.93. The associated RMSE was 2.6m/s without nudging and 1.2m/s with nudging.

Figure 2 shows the observed stability distribution at Shell Flats Mast 2 based on the Bulk Richardson number.


Figure 2: Observed stability distribution at Shell Flats Mast 2.

Figure 3 shows the same distribution inferred from the model data.


Figure 3: WRF model predicted stability distribution at Shell Flats Mast 2.

It can be seen that there is some difference in the distributions; the observed distribution shows a significant fraction of non-neutral events but the distribution shows a peak around neutral stability. In the case of the model, there is a clustering of the distribution at the (non-neutral) stability extremes, with a trough around neutral stability. Results are more encouraging when comparing observed and predicted stability trends diurnally and by season. Figure 4 shows a comparison between observed and predicted stability distributions at Shell Flats Mast 2 by month.


Figure 4: Comparison between observed and predicted stability distribution at Shell Flats Mast 2 by month.


It can be seen that the model captures the changing trend by month in stability distributions, even though the absolute distribution in each month is not captured as well.

Finally, an assessment was made of synoptic variability in the wind conditions across a large hypothetical Round 3 offshore wind farm in Dogger Bank. The offshore wind farm location has an extent of ~20km in each direction. Figure 4 shows an example of the maximum variation in wind direction across the wind farm due to prevailing synoptic conditions.


Figure 5: Predicted maximum wind direction difference across hypothetical offshore wind site extending ~20km in each dimension.

It can be seen that there is a difference ~50 degrees in wind direction across the extremities of the wind farm.


Conclusion

This research establishes the accuracy of a mesoscale model in assessing offshore wind resource for two relatively near offshore wind farm sites in UK waters. It has been seen that: 1) wind speeds could be predicted within an RMSE of less than 2m/s but this was site dependent; 2) the wind rose at an offshore site can be predicted with a reasonable accuracy, though localised directional turning effects were not captured; 3) model nudging provided clear benefits and suggested that offshore mast data could be extrapolated to a wider area with a reasonable degree of accuracy using a mesoscale model; 4) the absolute distribution of stability conditions was not predicted very accurately, but the relative change in conditions diurnally and by season was captured quite well; 5) there can be significant turning of the wind due to synoptic conditions across the extent of a large offshore wind farm which should be considered in predicting dynamic wake losses in conjunction with smaller scale turbulent wake meandering effects.
Prediction accuracy using a mesoscale model is suggested to be around 1.5-2m/s in terms of RMSE. It should be noted that this work has only considered two UK offshore sites at three masts and significant further work is required to establish levels of prediction accuracy. However, the results do suggest that the prediction of the offshore resource is more accurate for sites further offshore. This highlights the importance of accurate modelling of land/sea effects for near offshore sites. More investigation of the use of model ensemble runs and the accuracy of PBL schemes particularly for the prediction of offshore stability is suggested.



Learning objectives
The objective of this work is establish how well a mesoscale model can predict the wind resource at an offshore wind farm site. The accuracy of the prediction of the wind speed, wind direction and atmospheric stability conditions is assessed.


References
[1] Skamarock, W.C., Klemp, J.B., Dudhia, J., Gill, D.O., Barker, D.M., Duda, M.G., Huang, X-Y., Wang, W. and Powers, J.G. 2008. A description of the Advanced Research WRF Version 3. NCAR/TN-475=STR, NCAR Technical Note, Mesoscale and Microscale Meteorology Division, National Center of Atmospheric Research, June 2008, 113 pp.
[2] Janjic, Z. I., 2003: A Nonhydrostatic Model Based on a New Approach. Met. Atmos. Phy., 82, 271-285.
[3] Saha, S.,et al., 2010: The NCEP Climate Forecast System Reanalysis. Bull. Amer. Meteor. Soc., 91, 1015–1057.
[4] Shimada, S. and Ohsawa, T. 2011 Accuracy and characteristics of offshore wind speeds simulated by WRF. SOLA. 7, 21-24.
[5] Kwun, J.H., Kim, Y-K., Seo, J-W., Jeong, J.H. and You, S.H. 2009. Sensitivity of MM5 and WRF mesoscale model predictions of surface winds in a typhoon to planetary boundary layer parameterizations. Nat Hazards 51, 63–77.
[6] Peña, A., Hahmann, A.N., Hasager, C.B., Bingöl, F., Karagali, I., Badger, J., Badger, M., and Clausen, N.E. 2011. South Baltic Wind Atlas. South Baltic Offshore Wind Energy Regions Project. ISBN: 978-87-550-3899-8 Risø-R-1775(EN), Risø DTU.
[7] Raubenheimer, B., Ralston, D.K., Elgar, S., Giffen, D., Signell, R.P. 2012. Observations and predictions of summertime winds on the Skagit tidal flats, Washington. Continental Shelf Research, 60, s13-s21.
[8] Nawri, N., Petersen, G.N., Björnsson, H. and Jónasson, K. 2012. Evaluation of WRF mesoscale model simulations of surface wind over Iceland. Report VI2012–010. Icelandic Meteorological Office, ISSN 1670-8261. http://www.vedur.is/media/2012_010_web.pdf [Accessed 16/09/2013].
[9] Liu, Z., Liu, S., hu, F., Ma, Y. and Liu, H. 2012. A comparison study of the simulation accuracy between WRF and MM5 in simulating local atmospheric circulations over Greater Beijing. Science China Earth Sciences, 55(3), 418-427.
[10] Hughes J. 2013. Mesoscale modelling of the UK offshore wind resource. PhD thesis, Loughborough University, UK.