Fusion of Wavelet Transform Features for Reliable Fault Detection within an Ocean Turbine MCM System  
Author Janell Duhaney

 
Co-Author(s) Taghi M. Khoshgoftaar; Randall Wald; Pierre P. Beaujean

 
Abstract Through a case study involving experimental data, we demonstrate how feature level fusion can enable more reliable fault detection in a condition monitoring system for an ocean turbine. This study revolves around analyzing and interpreting vibration signals gathered from multiple accelerometers installed on various components of a dynamometer designed to test the drive train and generator of the turbine. We applied feature level fusion to combine these vibration readings, and then assessed the abilities of six well known machine learners to detect changes in state from the raw accelerometer data and from the fused data. Analysis of the performance of these classifiers showed more stable performances for the six classifiers in detecting the state of the machine from the fused data versus from the data from the individual sensor channels.

 
Keywords Ocean turbine, condition monitoring, reliability, feature level fusion, wavelet transform
   
    Article #:  1815
 
Proceedings of the 18th ISSAT International Conference on Reliability and Quality in Design
July 26-28, 2012 - Boston, Massachusetts, U.S.A.