基于SE-UniFormer的滚动轴承跨工况故障诊断方法
王冰

重庆三峡科技大学机械工程学院，重庆，404100；

摘要：针对滚动轴承在不同负载工况下振动信号分布差异明显、模型跨工况泛化能力不足的问题，提出一种基于 SE-UniFormer 的轴承故障诊断方法。首先，对振动信号样本进行快速傅里叶变换，获得频域幅值特征，并进行归一化处理；随后，将一维频域特征重排为二维特征图，输入 UniFormer 网络提取局部卷积特征和全局自注意力特征；最后，在 UniFormer 高层特征后引入 SENet 通道注意力模块，对不同通道特征进行自适应重标定，并通过全局平均池化和全连接层完成故障分类。以 CWRU 轴承数据集为对象，设置四组跨负载工况诊断任务进行验证。实验结果显示，所提方法在四组任务上的诊断准确率分别为 94.09%、100.00%、99.56% 和 91.82%，平均准确率达到 96.37%。这说明该方法在仅采用 FFT 频域特征作为输入的条件下，仍能获得较好的跨工况故障诊断效果，可为变负载条件下轴承故障识别提供一种有效方法。

关键词：滚动轴承；故障诊断；跨工况；快速傅里叶变换；UniFormer；SENet

参考文献

[1] ZHANG Z C, LI J, CAI C Z, et al. Bearing Fault Diagnosis Based on Image Information Fusion and Vision Transformer Transfer Learning Model[J]. Applied Sciences, 2024, 14(7): 2706. DOI: 10.3390/app14072706.

[2] ZHOU Q, MA W G, ZHANG Y D, et al. Bearing fault diagnosis for variable working conditions via lightweight transformer and homogeneous generalized contrastive learning with inter-class repulsive discriminant[J]. Engineering Applications of Artificial Intelligence, 2025, 139: 109548. DOI: 10.1016/j.engappai.2024.109548.

[3] BIAN C X, JIA C N, LI J B, et al. Rolling bearing fault diagnosis under small sample conditions based on WDCNN-BiLSTM Siamese network[J]. Scientific Reports, 2025, 15: 29591. DOI: 10.1038/s41598-025-12370-3.

[4] WEI L P, PENG X Y, CAO Y P. Multi-Scale Graph Transformer for Rolling Bearing Fault Diagnosis[J]. Eksploatacja i Niezawodnosc–Maintenance and Reliability, 2025, 27(2): 194779. DOI: 10.17531/ein/194779.

[5] LI K C, WANG Y L, ZHANG J H, et al. UniFormer: Unifying Convolution and Self-Attention for Visual Recognition[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2023, 45(10): 12581-12600. DOI: 10.1109/TPAMI.2023.3282631.

[6] HU J, SHEN L, ALBANIE S, et al. Squeeze-and-Excitation Networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020, 42(8): 2011-2023. DOI: 10.1109/TPAMI.2019.2913372.

[7] Case Western Reserve University. Bearing Data Center[EB/OL]. https://engineering.case.edu/bearingdatacenter.