感應馬達因其簡單的設計、穩健的操作和高效的能源利用，已經成為各種工業作業的基礎。近年來，基於深度學習的技術在偵測和診斷感應馬達中發生的各種異常和故障方面表現出色。然而傳統的深度學習架構也有其局限。例如，卷積神經網路（CNN）常常無法捕捉到數據關鍵特性的長期時間依賴性；同時遞迴神經網路（RNN）也有其缺點，特別是在處理長和複雜的數據序列時，容易受到梯度消失問題影響。為了克服這些挑戰，本研究開發了一個創新的深度學習框架，將一維卷積神經網路（1D-CNN）和遞迴神經網路（RNN）整合到兩個獨立的運算管道中。這個獨特的設定允許從原始信號輸入中直接提取空間特性以及時間、長期和短期依賴性。藉由加入一種原先為自然語言處理（NLP）所開發的多頭注意力機制，這個模型能夠更加重視相關的特性，同時降低過度擬合的風險。為了解決數據不足的問題，本論文也將生成對抗網路（GAN）融入到深度學習故障診斷框架中。GAN擅長產生與原始數據集相似的合成數據，因此能夠豐富訓練用的數據集。這解決了數據不足的基本問題，使模型能做出更準確和可靠的診斷。我們的方法在診斷感應馬達故障方面，與其他多種先進技術相比，都展現出有競爭力的準確度。本研究強調，注意力機制在提升為故障診斷而設計的深度學習模型性能方面扮演關鍵角色。透過允許直接從原始信號數據進行故障診斷，我們提出的方法有潛力顯著減少工業環境中的運行停機時間和相關成本。

Induction motors serve as the backbone of various industrial operations due to their inherent advantages, such as design simplicity, operation robustness, and energy efficiency. Over the past few years, techniques grounded in deep learning have demonstrated remarkable capabilities in detecting and diagnosing various anomalies and faults occurring in induction motors. However, conventional deep learning architectures have their limitations. For instance, Convolutional Neural Networks (CNNs) often fail to capture long-range temporal dependencies within the crucial features of the data. Concurrently, Recursive Neural Networks (RNNs) have their own drawbacks, such as susceptibility to the vanishing gradient problem and reduced computational efficiency, particularly when dealing with long and intricate data sequences. To address these challenges, this research introduces an innovative deep learning framework that seamlessly integrates a one-dimensional Convolutional Neural Network (1D-CNN) and a Recursive Neural Network (RNN) into two discrete computational pipelines. This unique configuration allows for extracting spatial features along with temporal, long-term, and short-term dependencies directly from the raw signal inputs. Incorporating a multi-head attention mechanism, a cutting-edge technique originally developed for Natural Language Processing (NLP), the proposed model can give higher importance to relevant features while mitigating the risk of overfitting. To address the issue of data scarcity, this thesis also incorporates a Generative Adversarial Network (GAN) into the deep learning fault diagnosis framework. GANs are adept at generating synthetic data that closely resembles the original dataset, thereby enriching the data pool available for training. This addresses the fundamental challenge of limited data availability, allowing the model to make more accurate and reliable diagnoses. Our method was found to offer a competitive level of accuracy in diagnosing faults in induction motors compared to several other state-of-the-art techniques. The study underscores the pivotal role played by attention mechanisms in enhancing the performance of deep learning models designed for fault diagnosis. By enabling the direct diagnosis of faults from raw signal data, the proposed method has the potential to dramatically minimize operational downtime and associated costs in industrial settings.

[1] A. H. Bonnett and G. C. Soukup, "Cause and analysis of stator and rotor failures in three-phase squirrel-cage induction motors," IEEE Transactions on Industry applications, vol. 28, no. 4, pp. 921-937, 1992.
[2] K. M. Siddiqui, K. Sahay, and V. Giri, "Health monitoring and fault diagnosis in induction motor-a review," International Journal of Advanced Research in Electrical, Electronics and Instrumentation Engineering, vol. 3, no. 1, pp. 6549-6565, 2014.
[3] C. Terron-Santiago, J. Martinez-Roman, R. Puche-Panadero, and A. Sapena-Bano, "A review of techniques used for induction machine fault modelling," Sensors, vol. 21, no. 14, p. 4855, 2021.
[4] X. Dai and Z. Gao, "From model, signal to knowledge: A data-driven perspective of fault detection and diagnosis," IEEE Transactions on Industrial Informatics, vol. 9, no. 4, pp. 2226-2238, 2013.
[5] P. Gangsar and R. Tiwari, "Signal based condition monitoring techniques for fault detection and diagnosis of induction motors: A state-of-the-art review," Mechanical systems and signal processing, vol. 144, p. 106908, 2020.
[6] Z. Gao, C. Cecati, and S. X. Ding, "A survey of fault diagnosis and fault-tolerant techniques—Part I: Fault diagnosis with model-based and signal-based approaches," IEEE transactions on industrial electronics, vol. 62, no. 6, pp. 3757-3767, 2015.
[7] T. Yang, H. Pen, Z. Wang, and C. S. Chang, "Feature knowledge based fault detection of induction motors through the analysis of stator current data," IEEE Transactions on Instrumentation and Measurement, vol. 65, no. 3, pp. 549-558, 2016.
[8] F. Filippetti, M. Martelli, G. Franceschini, and C. Tassoni, "Development of expert system knowledge base to on-line diagnosis of rotor electrical faults of induction motors," in Conference Record of the 1992 IEEE Industry Applications Society Annual Meeting, 1992: IEEE, pp. 92-99.
[9] M. E. H. Benbouzid, "A review of induction motors signature analysis as a medium for faults detection," IEEE transactions on industrial electronics, vol. 47, no. 5, pp. 984-993, 2000.
[10] Z. Wang, C. Chang, and Y. Zhang, "A feature based frequency domain analysis algorithm for fault detection of induction motors," in 2011 6th IEEE Conference on Industrial Electronics and Applications, 2011: IEEE, pp. 27-32.
[11] C. da Costa, M. Kashiwagi, and M. H. Mathias, "Rotor failure detection of induction motors by wavelet transform and Fourier transform in non-stationary condition," Case Studies in Mechanical Systems and Signal Processing, vol. 1, pp. 15-26, 2015.
[12] N. Bessous, S. Zouzou, W. Bentrah, S. Sbaa, and M. Sahraoui, "Diagnosis of bearing defects in induction motors using discrete wavelet transform," International Journal of System Assurance Engineering and Management, vol. 9, no. 2, pp. 335-343, 2018.
[13] J. Antonino-Daviu, P. J. Rodriguez, M. Riera-Guasp, A. Arkkio, J. Roger-Folch, and R. Pérez, "Transient detection of eccentricity-related components in induction motors through the Hilbert–Huang Transform," Energy Conversion and Management, vol. 50, no. 7, pp. 1810-1820, 2009.
[14] E. Sejdić, I. Djurović, and L. Stanković, "Fractional Fourier transform as a signal processing tool: An overview of recent developments," Signal Processing, vol. 91, no. 6, pp. 1351-1369, 2011.
[15] B. A. Boukamp, "Fourier transform distribution function of relaxation times; application and limitations," Electrochimica acta, vol. 154, pp. 35-46, 2015.
[16] R. J. Shakshi and R. Jaswal, "Brain wave classification and feature extraction of EEG signal by using FFT on lab view," Int. Res. J. Eng. Technol, vol. 3, pp. 1208-1212, 2016.
[17] K. S. Gaeid and H. W. Ping, "Wavelet fault diagnosis and tolerant of induction motor: A review," International Journal of the Physical Sciences, vol. 6, no. 3, pp. 358-376, 2011.
[18] G. R. Schleder, A. C. Padilha, C. M. Acosta, M. Costa, and A. Fazzio, "From DFT to machine learning: recent approaches to materials science–a review," Journal of Physics: Materials, vol. 2, no. 3, p. 032001, 2019.
[19] A. Choudhary, D. Goyal, and S. S. Letha, "Infrared thermography-based fault diagnosis of induction motor bearings using machine learning," IEEE Sensors Journal, vol. 21, no. 2, pp. 1727-1734, 2020.
[20] T. Han, C. Liu, W. Yang, and D. Jiang, "Learning transferable features in deep convolutional neural networks for diagnosing unseen machine conditions," ISA transactions, vol. 93, pp. 341-353, 2019.
[21] S. BHANDARI, "MECHANICAL FAULT ANALYSIS AND DETECTION USING NON STATIONARY DECOMPOSITION FOR VIBRATION SIGNALS," 2023.
[22] A. Apponi et al., "Heisenberg’s uncertainty principle in the PTOLEMY project: A theory update," Physical Review D, vol. 106, no. 5, p. 053002, 2022.
[23] S. Halder, S. Bhat, D. Zychma, and P. Sowa, "Broken rotor bar fault diagnosis techniques based on motor current signature analysis for induction motor—A review," Energies, vol. 15, no. 22, p. 8569, 2022.
[24] A. Almounajjed, A. K. Sahoo, and M. K. Kumar, "Condition monitoring and fault detection of induction motor based on wavelet denoising with ensemble learning," Electrical Engineering, vol. 104, no. 5, pp. 2859-2877, 2022.
[25] K. KaliappanSivakumar, G. Ganesan, G. S. Chermakani, and D. Muthukumar, "Identification of bearing fault in induction motor using random forest algorithm," in E3S Web of Conferences, 2023, vol. 387: EDP Sciences, p. 01005.
[26] M.-C. Kim, J.-H. Lee, D.-H. Wang, and I.-S. Lee, "Induction Motor Fault Diagnosis Using Support Vector Machine, Neural Networks, and Boosting Methods," Sensors, vol. 23, no. 5, p. 2585, 2023.
[27] M. Q. Tran, M. Amer, A. Y. Abdelaziz, H.-J. Dai, M.-K. Liu, and M. Elsisi, "Robust fault recognition and correction scheme for induction motors using an effective IoT with deep learning approach," Measurement, vol. 207, p. 112398, 2023.
[28] S. Chikkam and S. Singh, "Condition monitoring and fault diagnosis of induction motor using DWT and ANN," Arabian Journal for Science and Engineering, vol. 48, no. 5, pp. 6237-6252, 2023.
[29] T.-T. Vo, M.-K. Liu, and C.-L. Hsieh, "Hybrid 1D CNN-RNN Network for Fault Diagnosis in Induction Motors Using Electrical Signals," in 2023 International Conference on System Science and Engineering (ICSSE), 2023: IEEE, pp. 212-217.
[30] Y.-M. Hsueh, V. R. Ittangihal, W.-B. Wu, H.-C. Chang, and C.-C. Kuo, "Fault diagnosis system for induction motors by CNN using empirical wavelet transform," Symmetry, vol. 11, no. 10, p. 1212, 2019.
[31] H. C. Cho, J. Knowles, M. S. Fadali, and K. S. Lee, "Fault detection and isolation of induction motors using recurrent neural networks and dynamic Bayesian modeling," IEEE Transactions on Control Systems Technology, vol. 18, no. 2, pp. 430-437, 2009.
[32] Y. LeCun, L. Bottou, Y. Bengio, and P. Haffner, "Gradient-based learning applied to document recognition," Proceedings of the IEEE, vol. 86, no. 11, pp. 2278-2324, 1998.
[33] A. Choudhary, R. K. Mishra, S. Fatima, and B. Panigrahi, "Multi-input CNN based vibro-acoustic fusion for accurate fault diagnosis of induction motor," Engineering Applications of Artificial Intelligence, vol. 120, p. 105872, 2023.
[34] S. Hochreiter and J. Schmidhuber, "Long short-term memory," Neural computation, vol. 9, no. 8, pp. 1735-1780, 1997.
[35] P. Dhruv and S. Naskar, "Image classification using convolutional neural network (CNN) and recurrent neural network (RNN): A review," Machine Learning and Information Processing: Proceedings of ICMLIP 2019, pp. 367-381, 2020.
[36] S. Bodapati, H. Bandarupally, R. N. Shaw, and A. Ghosh, "Comparison and analysis of RNN-LSTMs and CNNs for social reviews classification," Advances in Applications of Data-Driven Computing, pp. 49-59, 2021.
[37] J. Wang, P. Fu, L. Zhang, R. X. Gao, and R. Zhao, "Multilevel information fusion for induction motor fault diagnosis," IEEE/ASME Transactions on Mechatronics, vol. 24, no. 5, pp. 2139-2150, 2019.
[38] J. Wang, P. Fu, S. Ji, Y. Li, and R. X. Gao, "A light weight multisensory fusion model for induction motor fault diagnosis," IEEE/ASME Transactions on Mechatronics, vol. 27, no. 6, pp. 4932-4941, 2022.
[39] L. Haque and M. J. Armstrong, "A survey of the machine interference problem," European Journal of Operational Research, vol. 179, no. 2, pp. 469-482, 2007.
[40] J. Gu et al., "Recent advances in convolutional neural networks," Pattern recognition, vol. 77, pp. 354-377, 2018.
[41] H. Gholamalinezhad and H. Khosravi, "Pooling methods in deep neural networks, a review," arXiv preprint arXiv:2009.07485, 2020.
[42] H. Habibi Aghdam, E. Jahani Heravi, H. Habibi Aghdam, and E. Jahani Heravi, "Convolutional neural networks," ed: Springer, 2017.
[43] H. Wang, Z. Liu, D. Peng, and Y. Qin, "Understanding and learning discriminant features based on multiattention 1DCNN for wheelset bearing fault diagnosis," IEEE Transactions on Industrial Informatics, vol. 16, no. 9, pp. 5735-5745, 2019.
[44] S. Huang, J. Tang, J. Dai, and Y. Wang, "Signal status recognition based on 1DCNN and its feature extraction mechanism analysis," Sensors, vol. 19, no. 9, p. 2018, 2019.
[45] Y. Hu, A. Huber, J. Anumula, and S.-C. Liu, "Overcoming the vanishing gradient problem in plain recurrent networks," arXiv preprint arXiv:1801.06105, 2018.
[46] K. Cho et al., "Learning phrase representations using RNN encoder-decoder for statistical machine translation," arXiv preprint arXiv:1406.1078, 2014.
[47] A. Mehrish, N. Majumder, R. Bharadwaj, R. Mihalcea, and S. Poria, "A review of deep learning techniques for speech processing," Information Fusion, p. 101869, 2023.
[48] A. Vaswani et al., "Attention is all you need," Advances in neural information processing systems, vol. 30, 2017.
[49] S. R. Joseph, H. Hlomani, K. Letsholo, F. Kaniwa, and K. Sedimo, "Natural language processing: A review," International Journal of Research in Engineering and Applied Sciences, vol. 6, no. 3, pp. 207-210, 2016.
[50] P. I. Frazier, "Bayesian optimization," in Recent advances in optimization and modeling of contemporary problems: Informs, 2018, pp. 255-278.
[51] Y. Jung, "Multiple predicting K-fold cross-validation for model selection," Journal of Nonparametric Statistics, vol. 30, no. 1, pp. 197-215, 2018.
[52] T. Takase, S. Oyama, and M. Kurihara, "Effective neural network training with adaptive learning rate based on training loss," Neural Networks, vol. 101, pp. 68-78, 2018.
[53] M. C. Cieslak, A. M. Castelfranco, V. Roncalli, P. H. Lenz, and D. K. Hartline, "t-Distributed Stochastic Neighbor Embedding (t-SNE): A tool for eco-physiological transcriptomic analysis," Marine genomics, vol. 51, p. 100723, 2020.
[54] P. Kumar and A. S. Hati, "Review on machine learning algorithm based fault detection in induction motors," Archives of Computational Methods in Engineering, vol. 28, pp. 1929-1940, 2021.
[55] I. H. Ozcan, O. C. Devecioglu, T. Ince, L. Eren, and M. Askar, "Enhanced bearing fault detection using multichannel, multilevel 1D CNN classifier," Electrical Engineering, vol. 104, no. 2, pp. 435-447, 2022.
[56] X. Wang, D. Mao, and X. Li, "Bearing fault diagnosis based on vibro-acoustic data fusion and 1D-CNN network," Measurement, vol. 173, p. 108518, 2021.
[57] D. Xiao, Y. Huang, X. Zhang, H. Shi, C. Liu, and Y. Li, "Fault diagnosis of asynchronous motors based on LSTM neural network," in 2018 prognostics and system health management conference (PHM-Chongqing), 2018: IEEE, pp. 540-545.
[58] L. Zhiwei, "Bearing Fault Diagnosis of End-to-End Model Design Based on 1DCNN-GRU Network," Discrete Dynamics in Nature and Society, vol. 2022, 2022.
[59] M. A. Bansal, D. R. Sharma, and D. M. Kathuria, "A systematic review on data scarcity problem in deep learning: solution and applications," ACM Computing Surveys (CSUR), vol. 54, no. 10s, pp. 1-29, 2022.
[60] I. Goodfellow et al., "Generative adversarial nets," Advances in neural information processing systems, vol. 27, 2014.
[61] Z. Zhang, M. Li, and J. Yu, "On the convergence and mode collapse of GAN," in SIGGRAPH Asia 2018 Technical Briefs, 2018, pp. 1-4.
[62] M. Arjovsky, S. Chintala, and L. Bottou, "Wasserstein generative adversarial networks," in International conference on machine learning, 2017: PMLR, pp. 214-223.
[63] B. Kamsu-Foguem, S. L. Msouobu Gueuwou, and C. A. K. A. Kounta, "Generative Adversarial Networks based on optimal transport: a survey," Artificial Intelligence Review, vol. 56, no. 7, pp. 6723-6773, 2023.
[64] I. Gulrajani, F. Ahmed, M. Arjovsky, V. Dumoulin, and A. C. Courville, "Improved training of wasserstein gans," Advances in neural information processing systems, vol. 30, 2017.
[65] A. Odena, C. Olah, and J. Shlens, "Conditional image synthesis with auxiliary classifier gans," in International conference on machine learning, 2017: PMLR, pp. 2642-2651.
[66] S. Cai, K. Zhu, Y. Ban, and T. Narumi, "Visual-tactile cross-modal data generation using residue-fusion gan with feature-matching and perceptual losses," IEEE Robotics and Automation Letters, vol. 6, no. 4, pp. 7525-7532, 2021.