Predicting the Performance of Offshore Wind Farm Using Artificial Intelligence

As the wind power industry continues to expand, grid presence of wind energy has significantly increased in the recent years. Short-term wind power predictions are becoming increasingly relevant because of the increasing penetration of wind power and the unpredictability in wind-electric generation caused by the fluctuating nature of wind. Physical approach, which combines Numerical Weather Predictions with wind farm performance models, are predominantly used in such forecasting systems. In this paper, the application of Artificial Intelligence in developing simple wind farm performance models, with the case of an offshore wind farm is demonstrated. Machine learning methods based on Artificial Neural Network, Support vector Machine, K- Nearest Neighbor and Random Forest are developed for predicting the power output from 40 turbines in the wind farm. With the minimum required inputs, these simplified models could perform well in estimating the wind farm performance.