基于K-means++与ELM的短期风电功率预测模型研究

The volatility of wind energy has brought great challenges to the rapid development of the wind power industry7. Accurate and reliable short-term wind power forecasting is essential to meet the requirement of power grid dispatching and reduce the cost per kilowatt hour of the electricity. This paper introduces a short-term wind power forecasting method based on /£-means ++ cluster analysis and ELM, meanwhile, the 72-hour wind power forecast is realized by using SCADA data and NWP data. Firstly, K-means ++ clustering algorithm is applied to divide the NWP data into clusters of varying numbers. Then, ELM model is used to establish a mapping model between \\\ V dala and SCADA power data for each cluster data. The best forecasting model is selected based on the distance between the data and the center point of each cluster after completing model training. The experimental results show that, compared with typical wind power forecast model, the proposed model has better performance in prediction accuracy.