• DocumentCode
    497353
  • Title

    Study on Engine Vibration Signal Denoising Method of SURE Block Threshold Based on DT-CWPT

  • Author

    Wu Dinghai ; Zhang Peilin ; Ren Guoquan ; Li Bing ; Cao Jianjun

  • Author_Institution
    Shijiazhuang Mech. Eng. Coll., Shijiazhuang, China
  • Volume
    1
  • fYear
    2009
  • fDate
    11-12 April 2009
  • Firstpage
    719
  • Lastpage
    722
  • Abstract
    A mechanical vibration signal denoise method of SURE block threshold based on DT-CWPT was proposed. The mechanical condition monitoring is always aims to removing the background noise and preserving the useful information more effectively. The presented method chooses the threshold level and block size by minimizing Steinpsilas unbiased risk estimate which is more adaptive and superior performance and can obtain higher signal-to-noise than the common NeighCoeff. Simultaneously, combine the method with the DT-CWPTpsilas characteristic of approximately shift-invariant and optimal representation for the signal. And the method was applied to denoise the engine vibration signal which has the property of complex frequency domain and nonstationary and time-variant. Testing results show that the proposed method can effectively compress the noise and the denoised signal has better classification performance.
  • Keywords
    acoustic signal processing; condition monitoring; engines; signal denoising; vibrations; SURE block threshold; engine vibration signal denoising method; mechanical condition monitoring; mechanical vibration signal demise method; Data mining; Discrete wavelet transforms; Engines; Frequency domain analysis; Noise reduction; Redundancy; Signal denoising; Signal processing; Vibrations; Wavelet transforms; block threshold; denoising; dual tree complex wavelet; vibration signal;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Measuring Technology and Mechatronics Automation, 2009. ICMTMA '09. International Conference on
  • Conference_Location
    Zhangjiajie, Hunan
  • Print_ISBN
    978-0-7695-3583-8
  • Type

    conf

  • DOI
    10.1109/ICMTMA.2009.529
  • Filename
    5203074