Int J Performability Eng ›› 2019, Vol. 15 ›› Issue (10): 2657-2666.doi: 10.23940/ijpe.19.10.p11.26572666

• Orginal Article • Previous Articles     Next Articles

Local and Global SR for Bearing Sensor-based Vibration Signal Classification

Shaohui Zhanga, Man Wanga, Canyi Dub,*, and Edgar Estupinanc   

  1. aSchool of Mechanical Engineering, Dongguan University of Technology, Dongguan, 523808, China bSchool of Automobile and Transportation Engineering, Guangdong Polytechnic Normal University, Guangzhou, 510665, China
    cDepartment of Mechanical Engineering, University of Tarapacá, Arica, Chile
  • Submitted on ; Revised on ; Accepted on
  • About author:* Corresponding author. <i>E-mail address</i>:
  • Supported by:
    This work was supported by the Natural Science Foundation of China (No 51605406, 51605405), Education and Scientific Research Project for Young and Middle-aged Teachers of Fujian (No JAT170413), Natural Science Foundation of Fujian Province (No 2018J01531), Research Start-up Funds of DGUT (No GC300501-26), and Postdoctoral Science Foundation of China (No 2019M652881)


Spectral regression (SR) is a method of feature extraction that realizes dimension reduction by the least squares method and can avoid eigen-decomposition of dense matrices. However, it only considers the affinity graph and misses the global information. In this paper, a novel feature extraction algorithm, called local and global spectral regression (LGSR), is proposed and applied to extract fault features from frequency-domain and time-domain features of vibration signals of bearing sensors. LGSR, which is the development of SR, is able to discover both local and global information of data manifold. Compared with other similar approaches (such as NPE, PCA, and SR), experiments of bearing defect classification validate that LGSR shows better ability to extract identity information for machine defect classification.

Key words: spectral regression, sensor-based signals, feature extraction, conditions classification