Short-term wind farm power forecasting with numerical weather prediction

Abstract

The deployment of wind farms in Brazil is beginning to affect the planning and scheduling of energy operation, mainly in south and northeast subsystems. As the wind is shifting, which can vary significantly even in a small interval of time, the System Operator must be prepared to act in the dispatch orders of the other sources. This paper presents a methodology for assessing meteorological and geographical data to measure the wind field in the turbine hub height, allowing to calculate the instantaneous electric power produced. This estimate can be extended to various points and times of day in order to monitor the variation of wind caused by climatic effects. The climatic variables are submitted to the California Meteorological Model (CALMET) to calculate wind fields in a three-dimensional mesh with a good assertiveness in the short term. Then, the forecast is used with the wind turbine model to obtain the electric power generated for a few hours ahead. With this result the System Operator can perform a more precise coordination of the hydrothermal short term dispatch, as well as the operative power reserve dimensioning.