点击次数：

论文类型：会议论文

发表时间：2010-01-01

收录刊物：EI、Scopus

卷号：5

页面范围：V5433-V5436

摘要：The resonance demodulation method can acquire the fault characteristic frequency of rolling element bearings, and then complete the accurate fault diagnosis of bearings. However, when rotating speed changes, due to the existence of effect of elastic hydrodynamic lubrication, the natural frequency and amplitude of bearing system will change with the speed changing, and the characteristic frequency no longer exists. This article discusses the order of nature of which the speed affect to the inherent frequency and the amplitude, establishes a fault vibration signal model of bearing in start-up process, extracts the fault feature signal by means of cyclic autocorrelation function, and then constructs AR coefficient characteristic vectors for CHMM model. Depending on the dynamics identifying ability of CHMM, we propose an effective fault diagnosis method, and it gets satisfactory results in the experiment. ? 2010 IEEE.