Rolling bearing fault identification method based on GAF-CNN-BiGRU network

Abstract

The invention discloses a rolling bearing fault identification method based on a GAF-CNN-BiGRU network. The method comprises the steps of firstly converting the vibration signal data of a rolling bearing into a two-dimensional image through a Gram angle field, and then completing the fault classification through a CNN-BiGRU network model, converting rolling bearing vibration signal data into a two-dimensional image by using the Gram angle field, so that complete information of an original signal is reserved, and dependence of the data on time is also reserved. In the CNN-BiGRU network model, spatial features in the two-dimensional image are extracted through a convolution unit, and time features of the two-dimensional image are further screened out through a bidirectional gate control unit, so that the accuracy of fault classification is improved.

Description

technical field [0001] The invention belongs to the technical field of fault identification of rotating machinery, and relates to fault identification of rolling bearings, in particular to a fault identification method of rolling bearings based on a GAF-CNN-BiGRU network. Background technique [0002] Rolling bearings are the core components of rotating machinery, and their health will directly affect the performance, stability and life of rotating machinery. Studies have shown that 40% to 50% of rotating machinery faults are related to the failure of rolling bearings. In order to ensure the safety of rotating machinery, it is of great significance to carry out effective fault diagnosis on rolling bearings. The intelligent fault diagnosis algorithm using machine learning has been widely used in the field of rolling bearing fault diagnosis. Although these methods have achieved good results, machine learning algorithms generally have a shallow structure, which limits the abil...

AI Technical Summary

Problems solved by technology

However, there are still some defects in the diagnosis process: the two-dimensional images converted from vibration signals are mostly grayscale images, which contain limited feature information; Complexity; in addition, the bearing vibration signal contains time dependence, especially for vibration signals with different fault degrees, this time relationship is particularly important

However, the traditional convolutional neural network is more about extracting local spatial features of images, and it is difficult to extract such time-dependent features, which affects the final fault identification accuracy of rolling bearings.

Images