轴承诊断中的频带细分与故障特征周期识别  

Frequency band division and fault characteristic period identification in bearing fault diagnosis

作　　者：彭伟[1,2] 杨圆鉴[1] 刘宇[1] 黄洪钟[1] 

机构地区：[1]电子科技大学机械电子工程学院,四川成都611731 [2]电子科技大学中山学院,广东中山528402

出　　处：《机械设计》2013年第11期38-41,共4页Journal of Machine Design

基　　金：国家自然科学基金资助项目(51075061);中央高校基本业务费资助项目(ZYGX2011J084);中山市科技计划资助项目(20123A338)

摘　　要：包络分析已经被广泛地应用于轴承的故障诊断中。其核心思想是利用一带通滤波器提高轴承故障信号的信噪比,然后利用包络解调分析提取低频轴承故障特征。谱峭度能有效地自动寻找轴承共振频率带,并用来提升轴承故障信号的信噪比,但由于其带通滤波后的高频轴承故障信号受到振动传递通道的影响,轴承故障特征时域分辨率能力仍较低。在高信噪比条件下,盲均衡算法能有效地去除传递通道的影响,提升轴承脉冲力时域分辨能力。因此,文中提出了一种新的轴承故障诊断算法,利用谱峭度和盲均衡算法提取轴承时域故障脉冲力。结果表明,该方法能有效地提高轴承故障时域检测能力。Envelope analysis has been widely used in bearing fault diagnosis.Its core idea is to use a band-pass filter to enhance the signal to noise ratio of bearing fault signal.Then,envelope demodulation analysis is employed to extract beating fault features.Spectral kurtosis is capable of automatically finding resonant frequency bands for the enhancement of the signal to noise of beating fault signal.However,the temporal discrimination of the beating fault signal obtained by spectral kurtosis analysis is still poor due to the influence caused by the vibration transmission channel.In the case of a high signal to noise ratio,blind equalization algorithm is effective in removing the influence from the vibration transmission channel and is able to improve the temporal discrimination of bearing fault signal.Considering the above two points,this paper reports a new beating fault diagnosis method,which consists of spectral kurtosis analysis and blind equalization algorithm.The results show that the developed method effectively enhances the temporal discrimination of beating fault signal.

关 键 词：包络分析 谱峭度 盲均衡 脉冲力 轴承诊断 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]