Abstract:Due to the shortage of traditional fossil energy and increasingly serious environmental pollution, green energy has become the development direction. Wind energy is a kind of clean and renewable energy, and the proportion of wind power generation in the national energy structure is enhancing increasingly. However, as the intermittent and volatility of wind power generation bring more and more serious impact on the power grid, the real-time power dispatch becomes particularly important. Related research works on ultra-short-term wind speed forecasting have drawn extensive attention. Firstly, the related domestic and international wind speed forecasting researches are comprehensively analyzed. Secondly, the research status of ultra-short-term wind speed forecasting methods are summarized from three aspects: physical methods, statistical and machine learning methods,and combination methods. Thirdly, the advantages and disadvantages of the main forecasting methods are compared.In addition, the uncertainty analysis and the error evaluation metrics are provided. Finally,the issues and directions that should be mainly addressed in the research of ultra-short-term wind speed forecasting are discussed.