An adversarial learning framework for zero-shot fault recognition of mechanical systems

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

18 Scopus citations

Abstract

Data imbalance is a major problem in intelligent fault diagnosis. Aiming at this problem, the paper proposed a novel adversarial learning framework for zero-shot fault recognition of mechanical systems. The proposed network consists of three parts which are the feature extractor, the generator and the discriminator. Trained with normal samples, the proposed method is capable of generating unseen fault samples by changing the condition of the generator. After, these synthetic samples are used to train an improved deep neural network for fault recognition. Results show that the proposed method can recognize the unseen faults even though none of fault samples are available during training, which is meaningful for industry application.

Original languageEnglish
Title of host publicationProceedings - 2019 IEEE 17th International Conference on Industrial Informatics, INDIN 2019
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages1275-1278
Number of pages4
ISBN (Electronic)9781728129273
DOIs
StatePublished - Jul 2019
Event17th IEEE International Conference on Industrial Informatics, INDIN 2019 - Helsinki-Espoo, Finland
Duration: 22 Jul 201925 Jul 2019

Publication series

NameIEEE International Conference on Industrial Informatics (INDIN)
Volume2019-July
ISSN (Print)1935-4576

Conference

Conference17th IEEE International Conference on Industrial Informatics, INDIN 2019
Country/TerritoryFinland
CityHelsinki-Espoo
Period22/07/1925/07/19

Keywords

  • Data imbalance
  • Deep learning
  • Fault detection
  • Rolling bearing

Fingerprint

Dive into the research topics of 'An adversarial learning framework for zero-shot fault recognition of mechanical systems'. Together they form a unique fingerprint.

Cite this