Short-Term Wind Power Forecasting by Advanced Machine Learning Models

Wind power forecasting receives more and more attention during the past decades. The intermittence and uncertainty of wind energy may cause unstable quality of power supply and bring great harm to industry. Therefore, a stable and accurate wind power forecasting technique is very important for wind energy system. A novel wind power prediction system is proposed in this paper. Key factors are firstly determined from the wind data provided by Central Weather Bureau of Taiwan. Then, serval machine and deep learning methods are adopted as backbone model to predict the power generation given a fixed interval. The performance is analyzed among different models in our experiment results.