TECHNICAL AREA Project: Installation of a model of prediction of the energetic production in Sotavento Wind Farm (LocalPred Model)

The prediction of energy produced in wind farms has a strategic nature in the actual situation of the electric market, because it is a phase of transition to free market. The prediction applied to wind farms constitutes a tool that will allow the management of the electric system in a more efficient way, providing capacity of predicition of wind power generation. The model of prediction allows the possibility of offers concerning the future energetic production of the farm.

The CIEMAT (Centre for Energetic, Environmental and Technological Researches) carried out, in the year 1999, a review of the the international state-of-the-art models of prediction for wind farms. CIEMAT has tested two models of prediction, the Prediktor of the Riso National Laboratory and the WPPT of the Technical University of Demmark in a wind farm of complex land in Spain.
The conclusions of the study have shown that:

• The models developed for flat land have not a good result on complex land.
• Weather forecasting in Spain needs a specific adaptation to the wind farm regarding local effects.

CIEMAT and CENER (National Centre for Renewable Energies) have stablished an agreement on the technical and commercial development of the techniques and programmes of wind prediction and wind power production.

Based on this agreement, Sotavento has carried out this project, together with CENER.

Sotavento Galicia, Ltd. as experimental wind farm, has showed a great interest on the valuation of prediction models, opting for the LocalPred model.

The main objectives are:

• Showing, in a practical way, the running of a model of prediction in a wind farm with complex land.
• Improving the integration of power produced ata the electric system, according to the regulations currently in force.
• Contributing to the daily planning of the electric request.

The different phases of the project are:

1. Adquisition of the numerical predictions.

It is necessary the acquisition of the numerical prediction of a meteorological model. In this case, we have chosen the National Institute of Meteorology (INM) and the HIRLAM model.
They will be used historical and real time predictions. Historical predictions will be useful in order to test the model of prediction and adapt it to the zone study.
2. Collection of wind data and production of the Farm.

Wind and energetic production data are necessary to the model of prediction. SOTAVENTO creates a database of measurements, adaptting it to the format required by the model of prediction.
3. Adjustment and optimization of the model of wind prediction.

The model of prediction was adjusted from the historical predictions and the average data of speed and production of the Farm.
4. Modelling of the Wind Farm running in relation to wind. The curve of the Farm.

The running of the wind farm will be modelled taking into account the data about the running and the operation of the farm supplied by SOTAVENTO, obtaining a power curve of the farm. For the modelling of the Farm they will be important the speed, wind direction, disponibility of the wind turbines and other parameters.
5. Reckoning of failures of the model of prediction.

The first reckoning of failures of the model will be carried out after the adjustment of the model to the wind farm.
6. Installation of the prediction model in Sotavento´s equipaments.

The installation of the model prediction will be carried out in SOTAVENTO´s equipments.
7. Evaluation of results.

Along the project, it will be carried out a tracking of the running of the model. The final phase of the project will consist in a evaluation of the output of the model of prediction.
8. Final phase in which the results obtained will be improved.

The results of the 7^th phase allow the identification of possible improvements of the model of prediction for Sotavento Wind Farm. The improvements are:
• More advanced statistic models.
• Physical modelling of high resolution with MM5 or FLUENT.

The present project of collaboration is carried out in the frame of the contacts maintained between SOTAVENTO, IDAE and CIEMAT with the purpose of installing a model of prediction of the energetic production in SOTAVENTO.

Nowadays, the project is finished.

At present, there are other short-term models of prediction (maximum extent: 120 hours). Some of these models are:

• MeteoLógica. It predicts power and provides a wind and direction map directed towards maintenance. It lasts for 92 hours.
• MeteoGalicia. Valuation of wind parameters (speed, direction, temperature, pressure and density). It lasts for 72 hours.