摘要
通过滚动轴承模拟故障试验台,获取了滚动轴承外圈、内圈和滚动体不同剥落程度时的振动信号,并对故障程度的识别与诊断进行了探索。采用经验模态分解方法对轴承信号进行分解,得到其固有模态分量,然后将前8阶分量的有效值作为特征向量输入BP神经网络,进行故障程度识别与诊断,滚动轴承3种类型不同程度的故障被准确地区分出来。
The vibration signals of the outer race, inner race and roller of bearing with various degrees of damages are collected at simulated fault test rig for roller bearing, identification and diagnosis of damage degree are studied. Bearing signals are decomposed into a series of intrinsic mode components by the method of empirical mode decomposition （ EMD）, and then the virtual values of former 8 components are put into BP artificial neural network as feature vectors to identify and diagnose the damage degree. Three kinds of bearing damages with different degree are identified correctly.
出处
《轴承》
北大核心
2008年第8期32-35,共4页
Bearing
关键词
滚动轴承
故障诊断
神经网络
经验模态分解
roller bearing
fault diagnosis
artificial neural networks
empirical mode decomposition
