State evaluation fusion index construction method with noise reduction effect

Abstract

The invention belongs to the technical field of electromechanical equipment state monitoring, and discloses a state evaluation fusion index construction method with a noise reduction effect, and the method comprises the following steps: S1, collecting data, and constructing a random matrix; s2, using ISSA to optimize oversampling parameters in RSVD, and performing noise reduction on the random matrix to obtain a noise-reduced random matrix; s3, performing RMT analysis on the random matrix after noise reduction, judging whether the current system is in a normal operation state or not, and constructing three LESs; s4, analyzing the weights of the three LESs by adopting an entropy method, and setting a dynamic threshold to screen the LES; and S5, performing KPCA fusion on the screened LES to obtain a state evaluation index. According to the method, the parameter selection of the RSVD is optimized by utilizing the ISSA, and the convergence efficiency of the algorithm is improved; a good noise reduction effect of the newly established LES is verified; and the entropy method is combined with the KPCA, so that the general applicability of fusing the LES as a state evaluation index to cope with different conditions of the system is improved.

Images