A fault diagnosis method based on semi-supervised learning deep adversarial network

Abstract

The invention discloses a fault diagnosis method based on a semi-supervised learning deep confrontation network, which obtains vibration signals of bearings under different working faults, and performs wavelet transformation on the vibration time domain signals of faulty bearings into two-dimensional images; Supervised learning is performed on labeled data, unsupervised training is performed on a large amount of unlabeled data, and convolutional neural networks are used to extract high-dimensional features to achieve data classification, thereby identifying bearing fault categories. The present invention realizes training to obtain a high-precision fault diagnosis model in the case of limited labeled data, and obtains a more accurate discriminator, thereby enabling accurate fault diagnosis based on the vibration signal of the rolling bearing.

Description

technical field [0001] The invention relates to the technical field of rolling bearing vibration signal processing, in particular to a fault diagnosis method based on a semi-supervised learning deep confrontation network. Background technique [0002] Researching advanced mechanical fault diagnosis methods is an important part of ensuring the safety of equipment and personnel. Among them, bearings are one of the most important mechanical parts in rotating machinery and are widely used in electric power, chemical industry, metallurgy, aviation and other important fields. One of the most easily damaged components, bearing performance and working conditions will directly affect the performance of the entire machine and equipment. Defects in bearing performance and working conditions will cause equipment to produce abnormal vibration and noise, and even cause equipment damage. Therefore, the fault diagnosis of rolling bearings, especially the analysis of early faults of rolling ...

AI Technical Summary

Problems solved by technology

However, using deep learning to train fault classification models requires a large amount of labeled sample data. However, in actual working conditions, although a large amount of vibration data can be collected, most of the vibration data do not have labels. For each fault It is unrealistic to collect a large amount of labeled vibration data

Therefore, the fault diagnosis method based on deep learning is not suitable for fault diagnosis of rolling bearings.

Images