基于机匣振动信号的滚动轴承故障特征提取

Fault Feature Extraction of Rolling Bearings Based on Casing Vibration Signals

通过进行带机匣测点的滚动轴承故障模拟实验,获取滚动轴承在故障状态条件下,轴承座测点和机匣测点的振动数据。分析结果显示,相对于轴承座,机匣上的振动信号成分复杂,轴承故障特征不明显,直接进行包络解调无法提取故障特征。通过奇异值分解(singular value decomposition,SVD),差分谱中各峰值处奇异值可以表征不同成分的信号。当轴承故障信号微弱时,第一个峰值处的奇异值重构信号往往代表转频及其调制信号分量,选取该靠后峰值处的奇异值进行信号重构可以有效提取轴承故障特征信号。研究内容为实际基于机匣测点信号的航空发动机滚动轴承故障特征提取提供了一种新的方法。

Fault feature extraction of rolling bearings based on casing vibration signal is studied. The experiment forthe fault simulation of the rolling bearings is done and the vibration signals of the bearing base and casing are acquired.Analysis results show that, compared to the bearing base, the vibration signal of the casing is complex and the fault featureof the bearing is not obvious. The envelope demodulation method cannot extract the fault characteristics directly. Therefore,the singular value decomposition (SVD) is employed to process the vibration signal. It is found that the singular values atdifferent peaks in the difference spectrum can represent the signals of different components. The singular valuereconstruction signal at the first peak always represents the components of the rotating frequency and modulation signalswhen the fault signals of the bearing are weak. The fault modulation signals can be effectively extracted by selecting thesingular values after the first peak in the difference spectrum. This study provides a new method for the fault featureextraction of rolling bearings based on the vibration signals of casing.