Abstract：

　　To improve the recognition of motor rotor imbalance faults, this paper presents a method based on the extreme point singular value decomposition (SVD) and the correlation dimension distribution. First, the motor vibration signals collected in different periods are de-noised using the extreme point singular value decomposition to avoid the misconvergence of correlation dimension caused by noise. Then the multiple correlation function replaces the traditional correlation function to determine the delay time, and a genetic programming (G-P) algorithm is utilized to calculate the correlation dimension. Finally, the motor rotor imbalance can be recognized by comparing the correlation dimension of multiple sets of vibration signals. We applied the developed method to a motor rotor imbalance in an oilfield in western China, and the results demonstrate that the correlation dimensions of de-noised signals can effectively identify the motor rotor imbalance. Compared with the traditional method, the correlation dimensions obtained by the developed method are more differentiable and more suitable for motor fault identification.

Key words： motor rotor imbalance fault feature extraction correlation dimension

Received： 15 November 2016

Corresponding Authors：jwang@cpu.edu.cn

Cite this article：YUAN Zhuang,DUAN Lixiang,WANG Jinjiang. Motor rotor imbalance fault recognition based on extreme point SVD de-noising and correlation dimension[J]. 石油科学通报, 2016, 1(3): 425-433.

URL：