Rolling bearing fault diagnosis system and method based on guiding type subfield self-adaption

Abstract

The invention provides a rolling bearing fault diagnosis system and method based on guiding type subfield self-adaption, and belongs to the bearing fault signal cross-domain intelligent fault diagnosis technology, and the system comprises a signal preprocessing module, a feedforward feature extraction network module and a subfield self-adaption module. According to the method, a source domain data set and a target domain data set are preprocessed and then inputted to a feedforward feature extraction network module, and more than one labeled guide sample under each type of fault is provided in the target domain; extracting signal features of the source domain data set and the target domain data set; and measuring the local distribution difference between the related sub-fields of the guide samples of the source domain and the target domain through the local maximum average difference, and minimizing the local distribution difference between the label-free samples of the source domain and the target domain to complete label prediction of the label-free samples of the target domain. According to the system and the method, cross-domain fault diagnosis among different types of bearing data sets can be realized by minimizing sub-domain differences among different types of bearing fault signals.

Description

technical field [0001] The invention relates to a bearing fault signal cross-domain intelligent fault diagnosis technology, in particular to a rolling bearing fault diagnosis system and method based on guided sub-field self-adaptation. Background technique [0002] As one of the essential parts in rotating machinery, bearings will pose a serious threat to the normal operation of the machinery once they fail. Reasonable analysis of bearing vibration signals is of great significance for early warning of mechanical failures and reduction of machine maintenance costs. With the continuous development of artificial intelligence technology, more and more scholars use intelligent network to classify rolling bearing fault types. The maturity of intelligent fault diagnosis technology is closely related to the development of deep learning. The powerful feature learning ability of the deep learning network makes fault diagnosis no longer rely on a large amount of professional knowledg...

AI Technical Summary

Problems solved by technology

However, when the feature gap between the target domain and the source domain is large, a large number of misjudgments will occur in the predicted labels, resulting in the forcible binding of subfields with different labels, which will reduce the accuracy of the classifier.

Images