Bearing residual life prediction method based on improved residual network and WGAN

Abstract

The invention relates to a bearing residual life prediction method based on an improved residual network and a WGAN, and the method comprises the steps: collecting original vibration signals of a bearing under different working conditions, and dividing the original vibration signals into a source domain signal and a target domain signal; sending the source domain signal and the target domain signal to an improved residual network to extract depth time sequence features; constructing a WGAN model, wherein the model comprises a feature generator and a field discriminator, using the improved residual network serves as the feature generator, the field discriminator discriminates the field from which the output feature of the feature generator comes, and if the field discriminator is difficult to distinguish, it is indicated that the feature generator successfully learns the common feature space of the sequences in the target domain and the source domain; and sending the common feature space to a full-connection neural network to construct a bearing residual life prediction model, and realizing prediction of the residual life of the target bearing. According to the invention, effective extraction and transfer learning of the vibration characteristics of the rolling bearing can be realized under variable working conditions and strong noise interference, so that accurate prediction of the residual life of the bearing is realized.

Description

technical field [0001] The invention relates to the technical field of equipment intelligent measurement and control, in particular to a method for predicting the remaining life of a bearing based on an improved residual network and a WGAN. Background technique [0002] Rolling bearings are important components in the field of mechanical transmission, and their performance often determines the service life of the entire mechanical system. Because of its importance and high failure rate, how to carry out fault diagnosis and remaining life prediction on rolling bearings has become a research hotspot in recent years. There are currently two mainstream RUL prediction methods, one is model-based and the other is data-driven. The model-based method needs to establish an extremely accurate mathematical or physical model to simulate the degradation trend of the bearing. Because of the high degree of difficulty in modeling, the data-driven method has become an important RUL method. ...

AI Technical Summary

Problems solved by technology

[0004] For this reason, the technical problem to be solved by the present invention is to overcome the problems existing in the bearing life prediction based on deep learning in the prior art. The present invention provides a bearing remaining life prediction based on the improved residual neural network and generation against WGAN, which can Realize effective extraction and transfer learning of rolling bearing vibration characteristics under variable working conditions and strong noise interference

Images