基于多尺度1DCNN的滚动轴承故障诊断

Fault Diagnosis of Rolling Bearing Based on Multi-scale 1DCNN

针对滚动轴承振动信号典型非平稳性、非线性的特点,提出一种基于小波变换(WT)和一维卷积神经网络(1DCNN)的轴承故障诊断多尺度卷积神经网络方法。通过小波变换对信号进行多尺度分解,然后对每个尺度成分进行重构,将重构后的信号进行傅里叶变换得到频谱表示,并将各尺度幅值数据构造成一维特征向量作为一维卷积神经网络的输入。最后利用一维卷积神经网络对输入数据进行特征学习,得到轴承故障诊断模型。利用滚动轴承的10个状态数据集验证其性能。结果表明:该方法可以避免人工提取特征,获得99.94%的诊断准确率。

Aiming at the typical non-stationary and non-linear characteristics of rolling bearing vibration signals,a multi-scale convolutional neural network method for bearing fault diagnosis based on wavelet transform and one-dimensional convolutional neural network was proposed.The signal was decomposed into multi scale components with wavelet transform,and then each scale component was reconstructed.The reconstructed signal was subjected to the Fourier transform to obtain the frequency spectrum representation,which was used as the input of the one-dimensional convolutional neural network.Finally,one-dimensional convolution neural network was used to learn the features of the input data and recognize the bearing fault.The performance of the model was verified by using 10 state data sets of rolling bearing.The results show that this method can avoid manual feature extraction and obtain a promising 99.94%diagnostic accuracy.