Rolling bearing fault diagnosis method based on SIFT-KPCA and SVM

Abstract

The invention provides a rolling bearing fault diagnosis method based on SIFT-KPCA and SVM. The rolling bearing fault diagnosis method comprises the steps of: firstly, converting vibration signals into a two-dimensional image, and utilizing wavelet denoising to reduce interference of noise on feature extraction before the conversion; secondly, extracting scale-invariant feature vectors of the two-dimensional image by adopting an SIFT algorithm to obtain a 128-dimensional feature matrix, and achieving dimension reduction of the feature vectors by adopting a KPCA algorithm; thirdly, and extracting singular values of the simplified feature vectors, and inputting the singular values into an SVM classifier to achieve fault classification finally. The rolling bearing fault diagnosis method has high classification accuracy.

Description

technical field [0001] The invention relates to the technical field of rolling bearing fault diagnosis, in particular to a rolling bearing fault diagnosis method based on SIFT-KPCA and SVM. Background technique [0002] Rolling bearings are widely used in rotating machinery and play a key role in maintaining the normal operation of equipment. Accidental bearing damage usually leads to serious mechanical failures and even huge economic losses. For this reason, it is particularly important to carry out accurate condition monitoring and fault diagnosis for rolling bearings. [0003] Bearing vibration signals contain rich system dynamic characteristic information, so vibration signal processing technology is one of the main tools for bearing fault diagnosis. Rolling bearing fault diagnosis mainly includes two key processes: feature extraction and pattern recognition. When a rolling bearing fails, its vibration signal becomes complex and non-linear, which makes effective featu...

AI Technical Summary

Problems solved by technology

However, the application of SIFT algorithm in the field of fault diagnosis research is still very little.

Images