Short-term wind power prediction method and system

Abstract

The invention discloses a short-term wind power prediction method and system. The method comprises the following steps: performing type identification, integer coding and spline interpolation on a multivariate time sequence, and uniformly processing a verification set and a test set by adopting standardized parameters calculated by a training set; evaluating the nonlinear correlation strength of each variable and the wind power by adopting a maximum information coefficient, and screening a key predictive factor according to an MIC score; decomposing the wind power sequence into a plurality of intrinsic mode functions, and suppressing high-frequency noise components during reconstruction to obtain a smooth target sequence; and inputting the processed data into a hybrid prediction model, extracting cross-feature interaction information through a 1 * 1 convolutional layer, capturing a time dependency relationship by using a Transform encoder, performing bidirectional sequence modeling through a BiLSTM network, and outputting a final wind power prediction value. According to the method, efficient denoising, feature optimization and multi-mode time sequence modeling of the wind power data are achieved, and the precision and stability of short-term wind power prediction are remarkably improved.

Images