在機械健康監測系統(Machine Health Monitoring System) 中，如何
判斷機械運作是否正常是常見且重要的問題。當機械加工時，同時由感測器收集刀具的加工訊號，並用此訊號來判斷加工是否出現異常，或是預測即將出現異常。傳統的物理模型或機器學習的方法，存在需大量人工標記、訓練或測試時間久、無法處理大量資料、標記成本高或不適用無加工異常的資料等問題。為了解決上述問題，我們提出一種用來預測加工訊號的框架 framework），包含三部分，地一部分為能加速訓練的資料前處理，第二部分為能準確預測刀具加工訊號的雙目標卷積類神經網路，第三部分為預測上下界，並使用此上下界作為加工狀態的參考。此框架中，不但不需人工標記，而且訓練網路的時間僅需將近 13分鐘。另外，在雙目標卷積類神經網路中，我們最終採用的是二維的卷積層，比起一維卷積層，能有效的提升預測的準確率。

In the Machine Health Monitoring System, how to judge whether the mechanical operation is normal is a common and important issue. When machining, the processing signal of the tool or machine is collected by the sensor, and this signal is used to judge whether the processing is abnormal or predict when an anomaly will occur. Traditional physical-based models or machine learning methods have problems such as requiring a large number of manual labels, training or testing for a long time, unsuitable for processing large amounts of data, high cost of labeling, or unable to process the data without fault case. In order to solve the above problems, we propose a framework for predicting CNC tool signal, which consists of three parts. The first part is the signal pre-processing that can accelerate the training networks. The second part is the Twin-Task Convolutional Neural Networks that can accurately predict the tool signals. The third part is to predict the upper and lower bounds, and we use this upper and lower bounds as a warning reference. In this framework, not only does it not require manual labeling, but the time of training networks takes only about 13 minutes. In addition, in the Twin-Task Convolutional Neural Networks, we finally use a two-dimensional convolutional layers, which can effectively improve the accuracy of predictions compared with the one-dimensional convolutional layers.

[1] Babu, Giduthuri Sateesh, Peilin Zhao, and Xiao-Li Li. "Deep convolutional neural network based regression approach for estimation of remaining useful life." International conference on database systems for advanced applications. Springer, Cham, 2016.
[2] Chandola, Varun, Arindam Banerjee, and Vipin Kumar. "Anomaly detection: A survey." ACM computing surveys (CSUR) 41.3 (2009): 15.
[3] Cheng, Fangzhou, et al. "Rotor-current-based fault diagnosis for DFIG wind turbine drivetrain gearboxes using frequency analysis and a deep classifier." IEEE Transactions on Industry Applications 54.2 (2017): 1062-1071.
[4] Ding, Xiaoxi, and Qingbo He. "Energy-fluctuated multiscale feature learning with deep convnet for intelligent spindle bearing fault diagnosis." IEEE Transactions on Instrumentation and Measurement 66.8 (2017): 1926-1935.
[5] Guo, Liang, et al. "Multifeatures fusion and nonlinear dimension reduction for intelligent bearing condition monitoring." Shock and Vibration 2016 (2016).
[6] Junbo, Tan, et al. "Fault diagnosis method study in roller bearing based on wavelet transform and stacked auto-encoder." The 27th Chinese Control and Decision Conference (2015 CCDC). IEEE, 2015.
[7] Liu, Ruonan, et al. "Dislocated time series convolutional neural architecture: An intelligent fault diagnosis approach for electric machine." IEEE Transactions on Industrial Informatics 13.3 (2016): 1310-1320.
[8] Malhotra, Pankaj, et al. "Multi-sensor prognostics using an unsupervised health index based on lstm encoder-decoder." arXiv preprint arXiv:1608.06154 (2016).
[9] Shao, Haidong, et al. "A novel deep autoencoder feature learning method for rotating machinery fault diagnosis." Mechanical Systems and Signal Processing 95 (2017): 187-204.
[10] Sohaib, Muhammad, Cheol-Hong Kim, and Jong-Myon Kim. "A hybrid feature model and deep-learning-based bearing fault diagnosis." Sensors 17.12 (2017): 2876.
[11] Sun, Wenjun, et al. “A sparse auto-encoder-based deep neural network approach for induction motor faults classification." Measurement 89 (2016): 171-178.
[12] Sun, Wenjun, et al. "Convolutional discriminative feature learning for induction motor fault diagnosis." IEEE Transactions on Industrial Informatics 13.3 (2017): 1350-1359.
[13] Verma, Nishchal K., et al. "Intelligent condition based monitoring of rotating machines using sparse auto-encoders." 2013 IEEE Conference on Prognostics and Health Management (PHM). IEEE, 2013.
[14] Verstockt, R. Van de Walle, S. Van Hoecke, Convolutional neural network based fault detection for rotating machinery, J. Sound Vib. 377 (2016) 331–345
[15] Verstraete, David, et al. "Deep learning enabled fault diagnosis using time-frequency image analysis of rolling element bearings." Shock and Vibration 2017 (2017).
[16] Wang, Peng, Ruqiang Yan, and Robert X. Gao. "Virtualization and deep recognition for system fault classification." Journal of Manufacturing Systems 44 (2017): 310-316.
[17] Wen, Long, et al. "A new convolutional neural network-based data-driven fault diagnosis method." IEEE Transactions on Industrial Electronics 65.7 (2017): 5990-5998.
[18] Yuan, Mei, Yuting Wu, and Li Lin. "Fault diagnosis and remaining useful life estimation of aero engine using LSTM neural network." 2016 IEEE International Conference on Aircraft Utility Systems (AUS). IEEE, 2016.[138] Zhao, Rui, et al. "Machine health monitoring with LSTM networks." 2016 10th International Conference on Sensing Technology (ICST). IEEE, 2016.
[19] Zhao, Rui, et al. "Deep learning and its applications to machine health monitoring." Mechanical Systems and Signal Processing115 (2019): 213-237.
[20] Zhao, Rui, et al. "Learning to monitor machine health with convolutional bi-directional LSTM networks." Sensors 17.2 (2017): 273.
[21] Zhao, Rui, et al. "Machine health monitoring with LSTM networks." 2016 10th International Conference on Sensing Technology (ICST). IEEE, 2016.