Model and Procedures for Reliable near Term Wind Energy Production Forecast

Abstract

Accurate and reliable prediction of wind energy production is important for the operational management of wind farm as well as ensuring the stability of electrical grid integrated with renewable wind energy. This paper describes a new method that aims to reliably predict wind energy production based on weather forecast data. With this new method, an aerodynamic model is firstly used to predict the distribution of wind speed with elevation by use of 24-hour ahead forecasted wind speed data at three-hour intervals (which is available with common weather forecast provider). The aerodynamic model considers the influence of factors such as ground topology, types of land cover, etc. on the wind speed distribution. Based on this model, wind speed at 10-minutes time interval is simulated, which is subsequently used together with the turbine production curve to predict energy production within the next 24 hours. The model and procedures for forecasting wind energy production are validated on a 100kW utility scale wind turbine. Comparison with alternative wind energy forecast procedures show that this new model-based forecast method provide more accurate prediction of wind energy production for different seasons. Major advantages of this model include that it is based on data commonly available from weather forecast providers and is applicable to different terrain conditions.