基于VMD-CNN的滚动轴承早期微弱故障智能诊断方法

Rolling Bearing Early Weak Fault Intelligent Diagnosis Based on VMD and CNN

针对滚动轴承早期微弱故障动力学响应弱、诊断分析严重依赖专业知识导致诊断难的问题,提出了一种基于变分模态分解分解与卷积神经网络的早期故障智能诊断方法。首先采用VMD对原始信号进行分解和高通滤波处理,消除低频高能量非故障信息的干扰,然后对滤波后高频信号进行包络和时频变换处理从而得到含有轴承早期故障信息的峰值能量谱,再将该峰值能量谱直接输入到卷积神经网络,利用卷积神经网络进行故障特征挖掘、提取和诊断模型的构建。最后采用所提方法对西安交通大学XJTU-SY滚动轴承故障数据进行分析,诊断准确率达97.0%以上,验证了所提方法对滚动轴承早期故障诊断的有效性。

To solve the problem of fault diagnosis caused by weak dynamic response of early fault of rolling bearing,an intelligent diagnosis method of early fault based on variational mode decomposition(VMD)and convolution neural network(CNN)is proposed.Firstly,VMD is used to decompose the original signal to eliminate the influence of low-frequency and non-fault information.Secondly,the envelope and time-frequency transformation of the filtered high-frequency signal are carried out to obtain the peak energy spectrum containing the bearing early fault information.Then,the peak energy spectrum is directly input to convolutional neural network for feature self-extraction and diagnosis model training.Finally,the fault data of XJTU-SY rolling bearing in Xi’an Jiaotong University are analyzed by using the proposed method,and the diagnosis accuracy is over 97.0%,which verifies the effectiveness of the proposed method for early fault diagnosis of rolling bearing.