Bearing fault diagnosis method based on deep adversarial migration network

Abstract

The invention discloses a bearing fault diagnosis method based on a deep adversarial migration network, which relates to the technical field of fault diagnosis of rotary machine vibration signals. Thediagnosis method comprises the steps: firstly, vibration frequency spectrum signals of a bearing under different working conditions (different rotating speeds and loads) are obtained, data with labels under a certain working condition serve as source domain data, and data without labels under other working conditions serve as target domain data; then a deep adversarial migration model based on agenerative adversarial neural network is then built to perform data training, two generators are adopted as a feature extraction network of a source domain and a target domain, Softmax cross entropy is adopted as a fault classifier, a discriminator is additionally adopted as a domain discrimination network, and a gradient inversion layer is added to perform domain discrimination training; and finally, network testing is carried out by adopting the residual target domain data. According to the method, the deep adversarial migration model is established by adopting the structure of the generative adversarial neural network, and fault feature migration learning and intelligent diagnosis of the bearing under different working conditions are efficiently and reliably realized.

Description

technical field [0001] The invention relates to the technical field of fault diagnosis of rotating machinery vibration signals, in particular to a bearing fault diagnosis method based on a deep anti-migration network. Background technique [0002] Mechanical equipment widely used in the fields of aerospace, rail transit, ocean engineering, and high-end CNC machine tools is developing towards high speed, automation, multi-function and precision, which makes the structure of mechanical equipment more and more complex, and the hidden danger of failure also follows. continue rising. The released "Mechanical Engineering Discipline Development Strategy Report (2011-2020)" has already listed the health status monitoring and fault diagnosis of complex mechanical systems as an important research direction. As a key component of mechanical power transmission, bearing has always been an important object of mechanical fault diagnosis. In order to carry out online fault monitoring on r...

AI Technical Summary

Problems solved by technology

However, the existing research is limited to the requirement that the difference between different distribution data is as small as possible, and the available data obtained under a certain working condition is sufficient, which is inconsistent with the characteristics of monitoring data of actual engineering equipment, and it is difficult to adapt to and meet the requirements of intelligent diagnosis of mechanical faults. Engineering Application Requirements

It is more difficult to accurately identify faults under the above-mentioned different working conditions, and it is also very difficult to label the newly obtained working condition data

Images