Wind power plant SCADA (Supervisory Control And Data Acquisition) data restoration method based on multiple correlation learning

The invention discloses a wind power plant SCADA data restoration method based on multi-correlation learning. The method comprises the following steps: S1, obtaining a wind power plant standardized data set # imgabs0 #; s2, obtaining a filling result # imgabs1 # under a global cross-correlation view angle; s3, obtaining a filling result # imgabS2 < 2 > under a global self-correlation view angle; s4, obtaining a filling result # imgabs3 # under a local cross-correlation view angle; s5, obtaining a filling result # imgabs4 # under a local autocorrelation view angle; s6, the four filling results of the # imgabs5 #, the # imgabs6 #, the # imgabs7 # and the # imgabs8 # are subjected to integration, denoising and reverse normalization through multiple linear regression, and a final repairing result # imgabs9 # is obtained. According to the method, autocorrelation and cross correlation inside SCADA multi-dimensional data are analyzed at the same time from the global and local aspects, and missing data are preliminar