Journal of Shanghai Jiaotong University ›› 2017, Vol. 51 ›› Issue (9): 1104-1110.

Previous Articles     Next Articles

 Research on Operation Condition Monitoring Method of Intellectual Apparatus Based on Data Driven

GAO Xiaoting,YANG Dongsheng    

  1.  College of Information Science and Engineering,Northeastern University
  • Online:2017-09-20 Published:2017-09-20
  • Supported by:
     

Abstract:  Alternating current (AC) contactor used frequently under full load is taken as a research object, and a method of apparatus operation condition monitoring based on data driven is proposed in this paper. First, historical operation data of AC contactor are collected by test platform and state characteristic parameters are gained. Combined with comprehensive assessment of wavelet transform and principal component analysis, data preprocessing such as denoising and removing outliers is used. Then, aimed at the disadvantage of highdimensionality and redundancy, kernel principal component analysis (KPCA) is adopted to merge multiinformation and kernel parameters are optimized based on test data. Finally, the fused information is input to hidden semiMarkov model (HSMM), and the operation condition monitoring and recognition of intellectual apparatus are realized. The practicability and validity of the method in apparatus condition monitoring is verified by test data from CJX28011 AC contactor.

Key words: kernel principal component analysis (KPCA), hidden semiMarkov model (HSMM), intellectual apparatus, wavelet transform,  condition monitoring, multiinformation fusion

CLC Number: