Short-term forecasting for multiple wind farms based on transformer model

• DOI: 10.1016/j.egyr.2022.02.184
• Publication Date: 2022-03-07T13:11:49Z
• Authors: Kai Qu; Gangquan Si; Zihan Shan; XiangGuang Kong; Xin Yang
• Keywords: Wind Power Forecasting; Probabilistic Forecasting

With the rapid growth of wind power installed capacity in recent years, distribution farms will be relatively dense, and there are usually multiple same area. However, due to complex correlations dependencies among these farms, traditional forecasting models for individual difficult apply. Meanwhile, accurate output is very important evaluation results renewable energy consumption grid, this problem has received extensive attention from many scholars. To improve accuracy forecasting, we apply Transformer model natural language processing (NLP) field forecasting. The proposed capable capturing longer sequence internal dependencies, as well key information data a comprehensive multifaceted way. By comparing with comparison method, case studies show that not only able accurately extract different levels correlation between but also give results.