eduzhai > Applied Sciences > Engineering >

FaultFace Deep Convolutional Generative Adversarial Network (DCGAN) based Ball-Bearing Failure Detection Method

  • king
  • (0) Download
  • 20210506
  • Save

... pages left unread,continue reading

Document pages: 36 pages

Abstract: Failure detection is employed in the industry to improve system performanceand reduce costs due to unexpected malfunction events. So, a good dataset ofthe system is desirable for designing an automated failure detection system.However, industrial process datasets are unbalanced and contain littleinformation about failure behavior due to the uniqueness of these events andthe high cost for running the system just to get information about theundesired behaviors. For this reason, performing correct training andvalidation of automated failure detection methods is challenging. This paperproposes a methodology called FaultFace for failure detection on Ball-Bearingjoints for rotational shafts using deep learning techniques to create balanceddatasets. The FaultFace methodology uses 2D representations of vibrationsignals denominated faceportraits obtained by time-frequency transformationtechniques. From the obtained faceportraits, a Deep Convolutional GenerativeAdversarial Network is employed to produce new faceportraits of the nominal andfailure behaviors to get a balanced dataset. A Convolutional Neural Network istrained for fault detection employing the balanced dataset. The FaultFacemethodology is compared with other deep learning techniques to evaluate itsperformance in for fault detection with unbalanced datasets. Obtained resultsshow that FaultFace methodology has a good performance for failure detectionfor unbalanced datasets.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×