基于VMD及模糊相关分类器的滚动轴承故障诊断

针对滚动轴承非平稳性的振动信号,提出一种基于变分模态分解(Variational Mode Decomposition,VMD)及模糊相关分类器的故障诊断方法。首先,对振动信号进行VMD分解,计算分解后分量与原信号的互信息值,利用互信息值提取无噪声分量,获得重构信号;其次,利用模糊函数在处理非平稳信号方面的优越性,结合相关系数提出模糊相关分类器;最后,将多组不同工作状态的重构信号输入模糊相关分类器,对多组数据进行训练与测试。实验结果表明,该方法能够有效的诊断出滚动轴承三种工作状态,且检测率较支持向量机及神经网络高。

Fault Diagnosis of Ball Bearing Based on VMD and ACC

Aiming at the no stationary characteristic of gear vibration signal, a recognition method based on Variational Mode Decomposition(VMD)and ambiguity correlation classification(ACC)is proposed. First, the vibration signal was decomposed by VMD, then a series of product function were obtained;Secondly, according to the mutual information, de-noised components was proposed to reconstruction, then, combining the correlation coefficient, ACC was proposed to identify the type of rolling bearing. Experimental results show that this method can effectively diagnosis three kinds of working condition, and the result is better than SVM or BP.