基于ITD和模糊聚类的齿轮箱故障诊断方法

为提高齿轮箱故障诊断的准确性与效率,针对其振动信号非线性和非平稳性的特点,提出将固有时间尺度分解(ITD)和模糊聚类(FCM)相结合的齿轮箱故障诊断方法。首先对齿轮箱振动信号进行固有时间尺度分解,提取包含主要故障信息的前4个固有旋转分量(PRC),求取PRC的特征能量作为故障特征向量。然后利用模糊C-均值聚类算法对齿轮箱故障进行识别与诊断,并将该方法应用到现场齿轮箱的诊断中。结果表明,诊断结果与实际情况完全相符,该方法比经验模式分解与模糊聚类相结合的方法具有更高的计算速度和精度,为齿轮箱故障诊断提供了一种新的有效方法。

Fault diagnosis method of gearbox based on intrinsic time-scale decomposition and fuzzy clustering

Considering the non-linear and non-stationary characteristics of vibration signals of gearbox, a new method based on intrinsic time-scale decomposition (ITD) and fuzzy center-means clustering (FCM) was proposed in order to diagnose gearbox ''s fault more accurately and effectively. Firstly, original vibration signals of gearbox were decomposed by ITD method. The first four proper rotation components (PRC) containing the main fault information were extracted and the PRC feature energy was calculated as fault feature vectors. Finally, faults of gearbox were identified by using FCM method. The results show that the diagnosis results of gearbox are totally in accordance with the actual situation in the application. The new method has high computation speed and accuracy compared with empirical mode decomposition (EMD), which provides a new efficient method for gearbox fault diagnosis.