|
|
|
| A New Fault Diagnosis Method Based on Local Mean Decomposition for Roller Bearings |
| XIE Ping1，YANG Yu-xin1，JIANG Guo-qian1,LI Xiao-li1，LI Xing-lin2 |
1.College of Electrical Engineering, Yanshan University, Qinhuangdao, Hebei 066004, China;
2. Hangzhou Bearing Test & Research Center, Hangzhou, Zhejiang 310022, China |
|
|
|
|
Abstract Targeting the characteristics of non-stationary and modulating of vibration signals for roller bearings, a new method for feature extraction based on Wigner-Ville spectral entropy is proposed. Firstly, the vibration signals are decomposed by the algorithm of local mean decomposition into several product functions. Secondly, the Wigner-Ville distribution of the principal components to obtain the time-frequency energy distributions is calculated, and Shannon entropy is introduced to construct a new index for feature extraction named Wigner-Ville spectral entropy. Finally, the feature vectors based on Wigner-Ville spectral entropy were input to least squares support vector machine, in order to automatically classify and diagnose the faults and damage degree of roller bearings. Simulation and experiments demonstrate the effectiveness and intelligence of the proposed method.
|
|
|
|
|
|
|
|
|
2014, Vol. 35 
