在現代工業中，旋轉機械是提供工廠穩定動力的來源之一，而其設備的穩定性是影響工廠生產及機器稼動率的重要因素。機械振動可能造成主軸偏心或軸承損壞，進而導致設備故障停機。近年來，深度學習模型被用來進行故障診斷，然而，也存在樣本數不平衡的問題以及不同的實驗機台無法通用模型等問題 。本研究提出基於深度遷移卷積神經網路模型對旋轉機械進行故障診斷建模，該框架是在深度CNN模型上構建的，輸入來自原始時域信號的預處理灰度圖像(64x64)。然後，將目標域遷移學習模型的全連接層替換為全局平均池化層，將不同資料集進行前處理後，對相似的信號特徵進行學習與歸納偏差，降低建模的時間與參數，並使模型對不同資料集具有更好的通用性。此模型使用開源資料集進行驗證，訓練集、驗證集與測試集分別為80%、10%與10%。經實驗結果顯示二個公開數據集轉移，並使用深度學習的指標準確率、混淆矩陣以及t-SNE降維可視化進行模型的驗證，再經由相似特徵的學習與歸納後 ，證明能夠有效解決樣本數不平衡的問題。其混淆矩陣的準確率在兩個資料集的相互遷移分別為CWRU資料集到Mendeley資料集為81.91%，Mendeley資料集到CWRU資料集為81.03%。而在VGG16與VGG19模型下，CWRU資料集的準確率分別為96.80%、99.79%，而Mendeley資料集的準確率分別為92.35%、95.12%，證明模型在遇到任何未知的資料時， 具有良好的通用性與適應性。

In modern industry, the stability of the equipment is an important factor in factory production and machine utilization. Mechanical vibration may cause spindle eccentricity or bearing damage in a rotating machine, which leads to equipment shutdown. In recent years, deep learning models are used for fault diagnosis because of their automatic feature learning capability. However, there are still some limitations such as unbalanced data problems which cannot be applied as a generalized model for a machine under different working conditions. This thesis proposes a deep transfer learning-based convolutional neural network framework for fault diagnosis of a rotating machine. Since transfer learning can solve similar but different problems efficiently by fine-tuning the knowledge learned from the source domain. The framework is made on a deep CNN model with the input of preprocessing grayscale images (64x64) from the original time domain signal. Then, the fully connected layer of the target domain transfer learning model is replaced with a global average pooling layer. The samples for the training set, validation set, and test set are 80%, 10%, and 10% respectively. With the testing by two similar open-source datasets, the deep learning accuracy indicators, confusion matrix, and t-SNE visualization are used to verify the reliability and robustness of the proposed algorithm. The results of transfer learning from CWRU to Mendeley dataset and from Mendeley to CWRU dataset achieve accuracy of 81.91%, and 81.03% respectively. With testing on two open-source models VGG16 and VGG19, the results on CWRU obtained 96.80%, and 99.79% accuracy respectively. The results on Mendeley also obtained 92.35%, and 95.12% accuracy respectively. It is provided that the proposed transfer learning algorithm is robust and could be used as a generalization model on similar machines.

[1] Y. Lei, B. Yang, X. Jiang, F. Jia, N. Li, and A. K. Nandi, "Applications of machine learning to machine fault diagnosis: A review and roadmap," Mechanical Systems and Signal Processing, vol. 138, p. 106587, 2020.
[2] J. Chen, C. Lin, D. Peng, and H. Ge, "Fault diagnosis of rotating machinery: a review and bibliometric analysis," IEEE Access, vol. 8, pp. 224985-225003, 2020.
[3] M. Xia, T. Li, L. Liu, L. Xu, S. Gao, and C. W. De Silva, "Remaining useful life prediction of rotating machinery using hierarchical deep neural network," in 2017 IEEE International Conference on Systems, Man, and Cybernetics (SMC), 2017: IEEE, pp. 2778-2783.
[4] A. L. Ellefsen, P. Han, X. Cheng, F. T. Holmeset, V. Æsøy, and H. Zhang, "Online fault detection in autonomous ferries: Using fault-type independent spectral anomaly detection," IEEE Transactions on instrumentation and measurement, vol. 69, no. 10, pp. 8216-8225, 2020.
[5] Y. Wang, H. Yang, X. Yuan, Y. A. Shardt, C. Yang, and W. Gui, "Deep learning for fault-relevant feature extraction and fault classification with stacked supervised auto-encoder," Journal of Process Control, vol. 92, pp. 79-89, 2020.
[6] A. He, & X. Jin., "Deep variational autoencoder classifier for intelligent fault diagnosis adaptive to unseen fault categories. " IEEE Transactions on Reliability, vol. 70, no. 4, pp. 1581-1595, 2021.
[7] S. Shen, Lu, M. Sadoughi, C. Hu, V. Nemani, A. Thelen, S. Kenny., "A physics-informed deep learning approach for bearing fault detection," Engineering Applications of Artificial Intelligence, vol. 103, p. 104295, 2021.
[8] W. Zhang, C. Li, G. Peng, Y. Chen, and Z. Zhang, "A deep convolutional neural network with new training methods for bearing fault diagnosis under noisy environment and different working load," Mechanical Systems and Signal Processing, vol. 100, pp. 439-453, 2018.
[9] J. Li, H. Wang, X. Wang, and Y. Zhang, "Rolling bearing fault diagnosis based on improved adaptive parameterless empirical wavelet transform and sparse denoising," Measurement, vol. 152, p. 107392, 2020.
[10] Y. Zhang, J. Ji, and B. Ma, "Fault diagnosis of reciprocating compressor using a novel ensemble empirical mode decomposition-convolutional deep belief network," Measurement, vol. 156, p. 107619, 2020.
[11] R. Dubey and D. Agrawal, "Bearing fault classification using ANN‐based Hilbert footprint analysis," IET Science, Measurement & Technology, vol. 9, no. 8, pp. 1016-1022, 2015.
[12] X. Yan and M. Jia, "A novel optimized SVM classification algorithm with multi-domain feature and its application to fault diagnosis of rolling bearing," Neurocomputing, vol. 313, pp. 47-64, 2018.
[13] R. Liu, F. Wang, B. Yang, and S. J. Qin, "Multiscale kernel based residual convolutional neural network for motor fault diagnosis under nonstationary conditions," IEEE Transactions on Industrial Informatics, vol. 16, no. 6, pp. 3797-3806, 2019.
[14] M. Sun, H. Wang, P. Liu, S. Huang, and P. Fan, "A sparse stacked denoising autoencoder with optimized transfer learning applied to the fault diagnosis of rolling bearings," Measurement, vol. 146, pp. 305-314, 2019.
[15] B. Ma, W. Cai, Y. Han, and G. Yu, "A novel probability confidence CNN model and its application in mechanical fault diagnosis," IEEE Transactions on Instrumentation and Measurement, vol. 70, pp. 1-11, 2021.
[16] S. Tang, S.Yuan, Y. Zhu, “Deep learning-based intelligent fault diagnosis methods toward rotating machinery”. IEEE Access, vol.8, pp. 9335-9346, 2019.
[17] S. Shao, R. Yan, Y. Lu, P. Wang, R. X. Gao. “DCNN-based multi-signal induction motor fault diagnosis.” IEEE Transactions on Instrumentation and Measurement, vol.69, no.6, pp.2658-2669. 2019.
[18] B. Zhao, C. Cheng, Z.Peng, Q. He, G. Meng. “Hybrid Pre-Training Strategy for Deep Denoising Neural Networks and Its Application in Machine Fault Diagnosis. ” IEEE Transactions on Instrumentation and Measurement, vol.70, pp.1-11. 2021.
[19] W. Zhang, C. Li, G. Peng, Y. Chen, Z. Zhang. “A deep convolutional neural network with new training methods for bearing fault diagnosis under noisy environment and different working load. ” Mechanical Systems and Signal Processing, no. 100, pp. 439-453, 2018.
[20] Z.-X. Hu, Y. Wang, M.-F. Ge, and J. Liu, "Data-driven fault diagnosis method based on compressed sensing and improved multiscale network," IEEE Transactions on Industrial Electronics, vol. 67, no. 4, pp. 3216-3225, 2019.
[21] L. Zhou, S. Pan, J. Wang, and A. V. Vasilakos, "Machine learning on big data: Opportunities and challenges," Neurocomputing, vol. 237, pp. 350-361, 2017.
[22] J. Shao, Z. Huang, and J. Zhu, "Transfer learning method based on adversarial domain adaption for bearing fault diagnosis," IEEE Access, vol. 8, pp. 119421-119430, 2020.
[23] J. Li, R. Huang, G. He, Y. Liao, Z. Wang, and W. Li, "A two-stage transfer adversarial network for intelligent fault diagnosis of rotating machinery with multiple new faults," IEEE/ASME Transactions on Mechatronics, vol. 26, no. 3, pp. 1591-1601, 2020.
[24] G. Xu, M. Liu, Z. Jiang, W. Shen, and C. Huang, "Online fault diagnosis method based on transfer convolutional neural networks," IEEE Transactions on Instrumentation and Measurement, vol. 69, no. 2, pp. 509-520, 2019.
[25] D. Neupane and J. Seok, "Bearing fault detection and diagnosis using case western reserve university dataset with deep learning approaches: A review," IEEE Access, vol. 8, pp. 93155-93178, 2020.
[26] H. Huang, N. Baddour. “Bearing vibration data collected under time-varying rotational speed conditions. ” Data in brief, vol. 21, pp.1745-1749. 2018.
[27] K. Simonyan, & A. Zisserman. “Very deep convolutional networks for large-scale image recognition. ” arXiv:1409.1556. 2014. (VGG16)
[28] L. Wen, X. Li, X. Li, and L. Gao, "A new transfer learning based on VGG-19 network for fault diagnosis," in 2019 IEEE 23rd international conference on computer supported cooperative work in design (CSCWD), 2019: IEEE, pp. 205-209.
[29] https://medium.datadriveninvestor.com/what-you-must-know-about-transfer-learning-4a6e4cb9fbad
[30] D. Zhang and T. Zhou, "Deep convolutional neural network using transfer learning for fault diagnosis," IEEE Access, vol. 9, pp. 43889-43897, 2021.
[31] Z. Chen, J. Cen, and J. Xiong, "Rolling bearing fault diagnosis using time-frequency analysis and deep transfer convolutional neural network," Ieee Access, vol. 8, pp. 150248-150261, 2020.
[32] Z. https://engineering.case.edu/bearingdatacenter/apparatus-and-procedures.
[33] https://data.mendeley.com/datasets/v43hmbwxpm/2
[34] Z. Zhu, Y. Lei, G. Qi, Y. Chai, N. Mazur, Y. An, and X. Huang, "A review of the application of deep learning in intelligent fault diagnosis of rotating machinery, " Measurement, vol. 206, Article. 112346, 2022.
[35] C. M. Tsai, C. S. Wang, Y. J. Chung, Y. D. Sun, and J. W. Perng, "Multisensor Fusion Time–Frequency Analysis of Thruster Blade Fault Diagnosis Based on Deep Learning, " IEEE Sensors Journal, vol. 22, no. 20, pp.19761-19771, 2022.
[36] J. Sun, J. Wen, C. Yuan, Z. Liu, and Q. Xiao, "Bearing Fault Diagnosis Based on Multiple Transformation Domain Fusion and Improved Residual Dense Networks, " IEEE Sensors Journal, vol. 22, no. 2, pp.1541-1551, 2021.
[37] D. Wang, Y. Li, L. Jia, Y. Song, and Y. Liu, "Novel three-stage feature fusion method of multimodal data for bearing fault diagnosis," IEEE Transactions on Instrumentation and Measurement, vol. 70, pp.1-10, 2021.
[38] Z. Huo, M. Martínez-García, Y. Zhang, and L. Shu, "A Multisensor Information Fusion Method for High-Reliability Fault Diagnosis of Rotating Machinery, " IEEE Transactions on Instrumentation and Measurement, vol. 71, pp.1-12, 2021.
[39] D. Wang, Y. Li, Y. Song, L. Jia, and T. Wen, "Bearing Fault Diagnosis Method Based on Complementary Feature Extraction and Fusion of Multisensor Data, " IEEE Transactions on Instrumentation and Measurement, vol. 71, pp.1-10, 2022.
[40] X. Nie, and G. Xie, "A two-stage semi-supervised learning framework for fault diagnosis of rotating machinery, " IEEE Transactions on Instrumentation and Measurement, vol. 70, pp.1-12, 2021.
[41] A. G. Nath, S. S. Udmale, D. Raghuwanshi, and S. K. Singh, "Improved structural rotor fault diagnosis using multi-sensor fuzzy recurrence plots and classifier fusion, " IEEE Sensors Journal, vol. 21, no.19, pp.21705-21717, 2021.
[42] A. G. Nath, S. S. Udmale, D. Raghuwanshi, and S. K. Singh, "Structural Rotor Fault Diagnosis Using Attention-Based Sensor Fusion and Transformers, " IEEE Sensors Journal, vol. 22, no. 1, pp.707-719, 2021.
[43] Y. Ma, and X. Wu, "Local-Dictionary Sparsity Discriminant Preserving Projections for Rotating Machinery Fault Diagnosis Based on Pre-Selected Multi-Domain Features, " IEEE Sensors Journal, vol. 22, no. 9, pp.8781-8794, 2022.
[44] F. Ferracuti, A. Freddi, A. Monteriu, and L. Romeo, "Fault diagnosis of rotating machinery based on wasserstein distance and feature selection, " IEEE Transactions on Automation Science and Engineering, vol.19, no. 3, 2021.
[45] Y. Wang, M. Yang, Y. Zhang, Z. Xu, J. Huang, and X. Fang, "A Bearing Fault Diagnosis Model Based on Deformable Atrous Convolution and Squeeze-and-Excitation Aggregation, " IEEE Transactions on Instrumentation and Measurement, vol. 70, pp.1-10, 2021.