Short-term Wind Power Forecast based on Teleconnected NWP Information

Accurate short-term wind power forecast (WPF) is essential for power system with high proportion of renewable energy (RE) integration. Teleconnected information of numerical weather predictions (NWPs) can give more abundant meteorological information for short-term WPF so as to improve the accuracy. A novel short-term WPF system considering teleconnected NWPs is established in this paper. Firstly, the relationship between teleconnected NWPs and wind power is analyzed quantitatively. Furthermore, a data cleaning (DC) strategy and WPF model structure combining teleconnected NWPs are design. We apply our algorithm based on actual operation data of a Shandong wind farm. Results show that after applying the framework built in this paper, the performance of common used short-term WPF models such as support vector regression (SVR), multi-layer perceptron (MLP) and long short-term memory (LSTM) have been improved effectively.