Bearing Fault Diagnosis Model Based on Multi-Level Domain Adaption Network

Abstract

The complex and changeable environment in the process of bearing operation may lead to inconsistent distribution of training data and test data, and decrease the diagnosis performance of the model. Thus a bearing fault diagnosis model based on the Shuffle-CANet is proposed, and realizes bearing cross-domain fault diagnosis by improving the ShuffleNet V2 and introducing asymmetric convolution. A domain loss function is added to the model based on the idea of domain adaptation in transfer learning so that the common features of the source domain and the target domain can be extracted occasionally and the cross-domain fault diagnosis can be realized. Compared with the traditional deep learning model, this model is friendlier to mobile and embedded devices. The Shuffle-CANet is validated by different transfer tasks on two different datasets. The results show that when the source domain and the target domain are derived from the same dataset, the fault diagnostic accuracy of the model can be more than 99%. When the target domain and the source domain are derived from different datasets, the fault diagnostic accuracy of the model can be more than 95%.

Keywords

bearing fault diagnosis / ShuffleNet V2 / multi-level domain adaption / lightweight

References

[1] YAN R Z, ZHANG Y B. Static and dynamic characteristics of vacuum preloaded porous aerostatic bearings[J]. Journal of Donghua University (English Edition), 2024, 41(1):81-88.

[2] WANG H C, XIANG G Q, GUO Z Q, et al. Fault diagnosis of rolling bearing compound faults based on improved time-frequency spectrum analysis method[J]. Journal of Aerospace Power, 2017, 32(7):1698-1703. (in Chinese)

[3] YANG X C. Rolling bearing fault diagnosis research based on wavelet spectrum analysis[J]. Coal Mine Machinery, 2013, 34 (1):289-291. (in Chinese)

[4] SARAVANAN N, RAMACHANDRAN K I. Incipient gear box fault diagnosis using discrete wavelet transform (DWT) for feature extraction and classification using artificial neural network (ANN)[J]. Expert Systems with Applications, 2010, 37(6):4168-4181.

[5] LI Y B, XU M Q, WEI Y, et al. A new rolling bearing fault diagnosis method based on multiscale permutation entropy and improved support vector machine based binary tree[J]. Measurement, 2016,77:80-94.

[6] SHAO J J, HUANG Z W, ZHU J M. Transfer learning method based on adversarial domain adaption for bearing fault diagnosis[J]. IEEE Access, 2020,8:119421-119430.

[7] HOANG D T, KANG H J. A bearing fault diagnosis method using transfer learning and Dempster-Shafer evidence theory[C]//Proceedings of the 2019 International Conference on Artificial Intelligence,Robotics and Control. New York: ACM, 2019:33-38.

[8] WANG Y, SUN X J, LI J, et al. Intelligent fault diagnosis with deep adversarial domain adaptation[J]. IEEE Transactions on Instrumentation and Measurement, 2021,70:1-9.

[9] HU X D, YANG X. Fault diagnosis method for rolling bearings based on siamese domain adversarial neural network[J]. Bearing, 2023(7):79-86. (in Chinese)

[10] CHEN R X, TANG L L, HU X L, et al. A rolling bearing fault diagnosis method based on deep attention transfer learning at different rotationals[J]. Journal of Vibration and Shock, 2022, 41(12):95-101,195.

[11] HUO C R, JIANG Q S, SHEN Y H, et al. Enhanced transfer learning method for rolling bearing fault diagnosis based on linear superposition network[J]. Engineering Applications of Artificial Intelligence, 2023,121:105970.

[12] MA N N, ZHANG X Y, ZHENG H T, et al. ShuffleNet V2:practical guidelines for efficient CNN architecture design[C]//15th European Conference. New York: ACM, 2018:122-138.

[13] HOU Q B, ZHOU D Q, FENG J S. Coordinate attention for efficient mobile network design[C]//2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). New York: IEEE, 2021:13713-13722.

[14] BISWAS K, KUMAR S, BANERJEE S, et al. SMU:smooth activation function for deep networks using smoothing maximum technique[EB/OL].(2022-04-11)[2023-06-02]. https://arxiv.org/abs/2111.04682v2.

[15] PAN S J, TSANG I W, KWOK J T, et al. Domain adaptation via transfer component analysis[J]. IEEE Transactions on Neural Networks, 2011, 22(2):199-210.

[16] ZHU Y C, ZHUANG F Z, WANG J D, et al. Deep subdomain adaptation network for image classification[J]. IEEE Transactions on Neural Networks and Learning Systems, 2021, 32(4):1713-1722.

[17] WANG C Y, MARK LIAO H Y, WU Y H, et al. CSPNet:a new backbone that can enhance learning capability of CNN[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW). New York: IEEE, 2020:390-391.

[18] SZEGEDY C, VANHOUCKE V, IOFFE S, et al. Rethinking the inception architecture for computer vision[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). New York: IEEE, 2016:2818-2826.

[19] LI Y W. Research on fault diagnosis method of rolling bearing under variable working condition based on improved CNN[D]. Harbin: Harbin University of Technology, 2019. (in Chinese)

[20] XIA Q, SHI T, WANG M Y. A level set based shape and topology optimization method for maximizing the simple or repeated first eigenvalue of structure vibration[J]. Structural and Multidisciplinary Optimization, 2011,43:473-485.

[21] DAGA A P, FASANA A, MARCHESIELLO S, et al. The Politecnico di Torino rolling bearing test rig:description and analysis of open access data[J]. Mechanical Systems and Signal Processing, 2019,120:252-273.