Ultra-short-term wind speed prediction method

The invention discloses an ultra-short-term wind speed prediction method, and relates to the technical field of wind speed prediction. The method comprises the following steps: acquiring wind measurement data of a wind power plant wind measurement tower, and performing quality control and filtering denoising on the wind measurement data to obtain a target wind measurement data sequence; dividing the target wind measurement data sequence into a wind speed modeling sequence and a test sequence; performing clustering processing on the wind speed modeling sequence to obtain different types of training data sets; and according to different categories of the training data set, respectively establishing a multi-step prediction model based on an extreme learning machine, performing multi-step wind speed prediction, and outputting a prediction sequence. According to the invention, multi-step wind speed prediction can be realized, and the prediction precision of wind measurement data is effectively improved. 本发明公开了一种超短