Entity

Time filter

Source Type

Chuzhou, China

Wang B.,Shanghai University | Wang B.,Chuzhou Institute | Liu S.-L.,Shanghai University | Zhang H.-L.,Shanghai University | Jiang C.,Shanghai University
Zhendong yu Chongji/Journal of Vibration and Shock | Year: 2015

Relevance vector machine (RVM) is a new statistical machine learning algorithm based on the theory of sparse Bayesian learning and it has been applied in many fields in recent years. The researchers both at home and abroad have paid more and more attentions to RVM. However, it has not received enough attention in the area of mechanical fault diagnosis. The characteristics of RVM were presented, its advantages and disadvantages were described compared with support vector machine (SVM). The domestic and overseas research advances, especially, those in the field of mechanical fault diagnosis were reviewed. Moreover, some existing problems in current research were analyzed, and the research directions of RVM for the mechanical fault diagnosis in the future were prospected. ©, 2015, Chinese Vibration Engineering Society. All right reserved. Source


Wang B.,Shanghai University | Wang B.,Chuzhou Institute | Liu S.-L.,Shanghai University | Jiang C.,Shanghai University | Zhang H.-L.,Shanghai University
Zhendong yu Chongji/Journal of Vibration and Shock | Year: 2015

A novel method to optimize relevance vector machine (RVM)'s kernel function parameters based on the quantum genetic algorithm (QGA) was proposed. It was compared with other optimization algorithms with simulations. The results showed that the optimization method based on QGA is superior to other optimization methods. The model of RVM optimized with QGA (QGA-RVM) was applied in fault diagnosis of rolling bearings. Fault signals were decomposed adaptively into some intrinsic mode functions (IMFs) with the ensemble empirical mode decomposition (EEMD). The IMF energy as fault features was inputted into QGA-RVM for final fault diagnosis. Experimental results showed that the proposed method can diagnose rolling bearings' faults rapidly and accurately, its validity and stability are verified; moreover, the superiority of RVM in intelligent fault diagnosis is revealed through the comparative analysis between QGA-RVM and SVM. ©, 2015, Chinese Vibration Engineering Society. All right reserved. Source


Zhan S.,Chuzhou Institute | Zhan S.,PLA Air Force Aviation University | Zhao S.,PLA Air Force Aviation University | Ni S.,Chuzhou Institute | And 3 more authors.
Qiangjiguang Yu Lizishu/High Power Laser and Particle Beams | Year: 2011

For the lack of precise control tools in spectral beam combining using reflecting volume Bragg grating, a simple experimental setup was designed using lenses as control elements of beam incident angle and collimation. The optimal grating parameters were derived based on the formula of diffraction efficiency, and the optimal lens parameters were derived based on the formula of aberrations. For the designed combining system, the influencing factors of combining results were analyzed, and the combining results for two laser beams with wavelengths of 1553 nm and 1559 nm were calculated. The results show that the combining efficiency of more than 90% can be achieved when the spectral widths of the two beams are less than 1.5 nm, and the combining efficiency of more than 88.7% can also be reached when the spectral widths of the two beams are equal to 2.1 nm. Source

Discover hidden collaborations