科研之友
微信
新浪微博
Facebook
分享链接
摘要
The fault signal of rolling element bearing in gearbox is weak and easy to be interfered，which leads to the difficulty of bearing fault feature extraction. In order to extract the weak fault feature in the signal, an adaptive multiscale generalized morphological filter ( AMGMF) was proposed. The morphological transformation can suppress interference and enhance the ability of feature extraction. Firstly, aiming at defects of the difference filter, a generalized enhanced difference filter ( GEDIF) was proposed. Amplitude frequency characteristics and pulse extraction characteristics were analyzed to reveal attributes of GEDIF. Secondly, lengths of flat structure elements were determined using local signal characteristics. Improved method of lengths selection could ensure adaptability and accuracy of morphological filter. Finally，weighted coefficients of structure elements were allocated with feature amplitude ratio ( FAR). The result of AMGMF was obtained after weighted reconstruction. Through analysis of simulated signal and planetary bearing fault signal, the results show that AMGMF can effectively separate fault features from complex signals. AMGMF has certain advantages, compared with single scale morphological filter, multiscale morphological filter and EEMD.
- 单位
