European experiences with wind resource assessment - An overview
By Kaj Jorgensen Risoe National Laboratory, Denmark

Introduction

For obvious reasons it is vital for any wind energy project to be based on reliable wind resource predictions. This is of significance both to make valid predictions of the payback on the wind power investments and to ensure the best integration into the power supply system. Hence, wind resource assessments may assist both in providing a sound basis for decisions on investments into wind energy (pre-investment) and in improving the operation of the power supply system to make the most of the wind energy (post-investment). The latter aspect is important from a systems approach, and at the same time such knowledge may enable the wind power operators obtaining better selling prices in deregulated power markets.

On the other hand, the needs with respect to the quality of the wind power predictions have to be offset against the opportunities in practice. First, there are restrictions on the time that may be spent on the exercise and usually restrictions that are not even close to the requirements to be able to extrapolate from the measurements with any degree of certainty. Secondly, costs are placing further constraints on the options.
Thus, basing the resource assessment on wind measurements carried out specifically for the project, is generally not a viable option. On the other hand, existing data may be available, even though these will usually need some for of adaptation to the specific sites in question.

Overview of approaches

On this background, a range of different methods exists that may be applied for the resource assessments. These methods vary from the most simplistic and uncertain, but low-cost, approaches on the one hand to highly sophisticated, but costly, on the other. Indeed, the different approaches may be combined to some extent at different stages of the project development.

Many approaches use the �regional wind climate� - that is wind data modified by removing impacts of local conditions - as point of departure. In other words, the regional wind climate illustrates geographical and temporal variations in wind statistics reduced to standard conditions, excluding the impact of roughness and near-by obstacles among others. To translate the regional wind climate into wind predictions for specific sites and hub heights, most approaches apply various types of microscale modelling. In the past, computing requirements have been prohibitive for the application of models of medium and large-scale weather events - mesoscale modelling - for this purpose. The computer development is rapidly reducing this obstacle and the coming years may see mesoscale modelling applied for wind resource predictions, either on its own or in combination with microscale modelling.

Among the most sophisticated approaches, are methods based on global databases containing data on wind, temperature and pressure in a global grid. For instance, the European Centre for Medium-range Weather Forecasting project has such a database. Potentially, such global databases are highly interesting, but for the time being their resolution is low.

Regional wind climate and microscale modelling - the Wind Atlas and the WAsP model

Risoe National Laboratory developed the Wind Atlas method for Denmark in the early 1980s and today this is the most widely used wind registration scheme in the world, having been applied in more than 90 countries. In the first step, the method establishes a generalised regional wind climate based on inputs regarding sheltering obstacles, roughness of the terrain and height contour lines. Secondly, a wind characterisation is generated of the specific site in point.

The Wind Atlas usually is generated by means of the microscale model Wind Analysis and Application Programme, WAsP, but other options exist. WAsP was developed by Risoe in 1993 as a computer tool for the practical application the wind atlas tool. {}

Short-term predictions

The objective of short-term prediction of the wind is to provide a better basis for the integration of fluctuating wind power supplies into the power supply system by predicted the wind power output 1-2 days in advance. This is particularly important in countries with high wind penetration levels, such as Denmark.

Short-term prediction usually based on the output from Numerical Weather Prediction models, operated by meteorological services. These typically provide 48 hours forecasts with runs every 6 to 12 hours. The NWP output is then combined with key features of the wind turbine or wind farm to generate predictions of the power output. In addition off-line or on-line data of measurements of its output are used for updating of the model parameters (by means of regression or autoregression).
Such short-term prediction models for wind power have been developed since the mid-1980s. Examples of models currently in use, include Prediktor, WPPT (Wind Power Prediction Tool) and MORE-CARE but a considerable number of others exist and also there is a constant rapid development in this field. In addition to improving the validity of the predictions within the 1-2 days� range of the present models, attempts are being made to increase the range of the forecasts (e.g. to 5 days). These improvements are being facilitated by the increase in opportunities offered by the ongoing computer development.

Conclusions

Wind resource predictions, both prior to investments and after, are of crucial importance to the success of wind energy projects. The methods in this field are in a state of rapid development, not least supported by the general computer development. Besides the general need for further improvement of the validity of the methods, there is a need for a range of approaches at different levels. In particular, methods for the early stage choice of the most suitable sites for wind power would be useful.

References

Ackermann, T. & L. Soder, 2002. An overview of wind energy - status 2002, Renewable and Sustainable Energy Reviews, Vol. 6, pp. 67-128.

Hansen, J. C., 2003. Experiences and recommendations on development of wind energy projects, ASEM Green IPP Network Workshop, Risoe National Laboratory, Roskilde, Denmark, 27 March

Landberg, L., 2001. Short-time prediction of local wind conditions, Journal of Wind Engineering and Industrial Aerodynamics, Vol. 89, pp. 235-45.

Landberg, L. et al, 2003a. Wind resource estimation - an overview, Wind Energy, Vol. 6, No. 3, pp. 261-71.

Landberg, L. et al, 2003b. Short-term prediction - an overview, Wind Energy, Vol. 6, No. 3, pp. 273-80.

Lange, B & J. H?jstrup, 2001. Evaluation of the wind-resource estimation program WAsP for offshore applications, Journal of Wind Engineering and Industrial Aerodynamics, Vol. 89, pp. 271-91.

Mortensen, N. G. et al, 1993. Wind Atlas Analysis and Application Program (WAsP), Roskilde, Denmark.

Starkov, A. N. et al, 2000, Russian Wind Atlas, Moscow Russia.
Troen, I. & E. L. Petersen,  1989. European Wind Atlas, Roskilde, Denmark.