eduzhai > Applied Sciences > Engineering >

Temporal clustering network for self-diagnosing faults from vibration measurements

  • Save

... pages left unread,continue reading

Document pages: 9 pages

Abstract: There is a need to build intelligence in operating machinery and use dataanalysis on monitored signals in order to quantify the health of the operatingsystem and self-diagnose any initiations of fault. Built-in control procedurescan automatically take corrective actions in order to avoid catastrophicfailure when a fault is diagnosed. This paper presents a Temporal ClusteringNetwork (TCN) capability for processing acceleration measurement(s) made on theoperating system (i.e. machinery foundation, machinery casing, etc.), or anyother type of temporal signals, and determine based on the monitored signalwhen a fault is at its onset. The new capability uses: one-dimensionalconvolutional neural networks (1D-CNN) for processing the measurements;unsupervised learning (i.e. no labeled signals from the different operatingconditions and no signals at pristine vs. damaged conditions are necessary fortraining the 1D-CNN); clustering (i.e. grouping signals in different clustersreflective of the operating conditions); and statistical analysis foridentifying fault signals that are not members of any of the clustersassociated with the pristine operating conditions. A case study demonstratingits operation is included in the paper. Finally topics for further research areidentified.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×