残差网络和注意力机制相结合的滚动轴承故障诊断模型

Rolling Bearing Fault Diagnosis Model Combining with Residual Network and Attention Mechanism

针对轴承故障诊断过程中存在的特征提取复杂、分类器训练困难等问题,提出了一种基于残差网络和注意力机制相结合的滚动轴承故障诊断模型。该模型以滚动轴承的一维振动时序信号作为输入,通过残差网络完成特征提取,然后经带有注意力机制的双向长短记忆神经网络单元,实现特征在时序上的表达并赋予不同的权重,输出到分类器完成端到端的振动信号分类,完成滚动轴承故障的诊断。实验表明,该模型的诊断准确率可达99.86%以上,对各故障类别的诊断率均在99%以上,提取的特征信息区分度高;模型诊断准确率优于基于特征工程的诊断模型,稳定性优于其他基于深度学习的诊断模型。

Aiming at the problem of feature extraction and classifier training in bearing fault diagnosis, a new fault diagnosis model of rolling bearing based on residual network and attention mechanism is proposed in this paper. The diagnosis model takes onedimensional vibration time series signal as its input, completes feature extraction through residual network, and then achieves feature expression in time domain through bi-directional long short-term meomory unit with attention mechanism, assigns different weights, and outputs them to classifier to complete end-to-end vibration signal diagnosis. Bearing fault diagnosis experiments show that the diagnostic accuracy of the proposed model is over 99.86%, and the diagnostic accuracy of each fault type is over 99%, and the extracted feature information is highly distinguished. The diagnostic accuracy of the model is superior to the diagnostic models based on feature engineering. Compared with the models based on deep learning, the proposed model has better stability.