Loading...

Fault Detection of Rotary System with Signal Processing and Intelligent Systems

Tebyanian, Afshin | 2011

740 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 41504 (08)
  4. University: Sharif University of Technology
  5. Department: Mechanical engineering
  6. Advisor(s): Behzad, Mehdi
  7. Abstract:
  8. Condition monitoring and fault detection is an important way to provide economy in both time and costs. Many methods are developed for this purpose. Vibration analysis is one of the most important ways in this field. In this thesis, fault detection of mechanical rotary systems with nonlinear and non-stationary nature has been developed by use of empirical mode decomposition (EMD) and Hilbert transform. Firstly the measured signal is transformed into some intrinsic mode functions (IMF’s) by EMD, and then the Hilbert transform is implemented on each IMF. With Hilbert transform, the instantaneous frequency of system and then the mean frequency are obtained. Mean frequency is a key definition for extracting special features for each signal. These features are applied for fault classification. At next step, mean frequency is reduced to some finite key coefficients by autoregressive method. These features are final features and introduced to a neural network for classification. Ball bearings of a rotary system and an industrial gas turbine are analyzed with this system. Results have been demonstrated that proposed fault detection system is able to detect faults of these faulty systems accurately
  9. Keywords:
  10. Fault Detection ; Hilbert Transform ; Autoregressive Process ; Neural Network ; Empirical Modes Decomposition (EMD)Method ; Midle Frequency

 Digital Object List

  • محتواي پايان نامه
  •   view

 Bookmark

No TOC