人工智慧的興起，使得訊號處理有突破性的發展，透過振動訊號的診斷，可以即時的反應刀具切削狀態。本研究根據刀具磨耗狀態進行監控，刀具壽命大致可分為三期:初始磨耗階段(Initial wear stage)、穩定磨耗階段(Steady wear stage)與劇烈磨耗階段(Severe wear stage)。在實驗上，本研究以兩種實驗方式，變切削參數實驗與定切削參數實驗，來探討訊號與刀具磨耗狀態關係。在變切削參數實驗中，透過振動訊號來還原工件表面影像，並利用該影像之表面紋路清晰度與原始工件影像對比，找出能夠反映刀具磨耗的頻寬訊息；在定切削參數實驗中，使用三種非監督式學習方法來評估，刀具在長時間銑削訊號的變化，並利用健康指數(Health index, HI)來反映訊號的衰退情形。此三種方法分別是:1.運用自編碼壓縮訊號2.運用自編碼之重構誤差3.運用卷積自編碼之重構誤差。根據HI的衰退趨勢，能夠了解特定頻寬的訊號與刀具磨耗有關，反映出來的頻寬與變切削參數實驗有相同的結果。從分析結果來看，主軸z方向之imf3訊號最能反映刀具磨耗的變化，此訊號能夠辨識出刀具即將進入劇烈磨耗階段之時刻，提供及時警訊。透過將imf6訊號進行格拉姆角場(Gramian Angular Field, GAF)轉換，使模型有更多的辨識特徵，經由自編碼分析後，能夠辨識出刀具一、二階段。最後利用多尺度熵分析，了解訊號發生不穩定的時刻點，對照刀具磨耗階段，發生在刀具初始及穩定磨耗階段交界處，並透過頻譜分析來驗證以上分析方法的準確性。

The rise of artificial intelligence has led to a breakthrough in signal processing. Through the diagnosis of vibration signals, the cutting state of the tool can be reflected in real time. In this study, the tool wear stages are monitored. Tool life can be roughly divided into three stages: initial wear stages, stable wear stages and severe wear stages. This study uses two experimental methods, variable cutting parameter experiment and fixed cutting parameter experiment, to explore the relationship between signal and tool wear stages.
In the variable cutting parameter experiment, the vibration signal is used to reconstruct the surface image of the workpiece, and the reconstructed surface image is compared with the actual image of the workpiece to find out the information of the frequency bandwidth that can reflect the tool wear. In the fixed cutting parameter experiment, three unsupervised learning methods are used to evaluate the changes in the signal during long-term milling of the tool, and we also use the Health Index (HI) to reflect the decline of the signal. The three methods are: 1. Use Autoencoder to compress the signal 2. Use reconstructtion error of Autoencoder. 3. Use reconstructtion error of Convolutional Autoencoder.
According to the declining trend of HI, it can be understood that the signal of a specific bandwidth is related to tool wear, and the reflected bandwidth has the same result as the experiment of variable cutting parameters. From the analysis results, the imf3 signal of the spindle can best reflect the changes in tool wear. This signal can identify the moment when the tool is about to enter the severe wear stage and provide timely warning. By converting imf6 signal to Gramian Angular Field (GAF), the model has more identification features. After Autoencoder analysis, the first and second stages of the tool can be identified. Finally, the multi-scale entropy analysis is used to understand the time when the signal is unstable, compared to the tool wear stage, which occurs at the junction of the initial and stable tool wear stage, and the accuracy of the above analysis method is verified through spectrum analysis.

[1] K. Salonitis and A. Kolios, "Reliability assessment of cutting tool life based on surrogate approximation methods," The International Journal of Advanced Manufacturing Technology, vol. 71, no. 5-8, pp. 1197-1208, 2014.
[2] J. Karandikar, T. McLeay, S. Turner, and T. Schmitz, "Tool wear monitoring using naive Bayes classifiers," The International Journal of Advanced Manufacturing Technology, vol. 77, no. 9-12, pp. 1613-1626, 2015.
[3] S. Dutta, A. Kanwat, S. Pal, and R. Sen, "Correlation study of tool flank wear with machined surface texture in end milling," Measurement, vol. 46, no. 10, pp. 4249-4260, 2013.
[4] Z. Hao, D. Gao, Y. Fan, and R. Han, "New observations on tool wear mechanism in dry machining Inconel718," International Journal of Machine Tools and Manufacture, vol. 51, no. 12, pp. 973-979, 2011.
[5] N. Ghosh et al., "Estimation of tool wear during CNC milling using neural network-based sensor fusion," Mechanical Systems and Signal Processing, vol. 21, no. 1, pp. 466-479, 2007.
[6] C. Drouillet, J. Karandikar, C. Nath, A.-C. Journeaux, M. El Mansori, and T. Kurfess, "Tool life predictions in milling using spindle power with the neural network technique," Journal of Manufacturing Processes, vol. 22, pp. 161-168, 2016.
[7] M. Nouri, B. K. Fussell, B. L. Ziniti, and E. Linder, "Real-time tool wear monitoring in milling using a cutting condition independent method," International Journal of Machine Tools and Manufacture, vol. 89, pp. 1-13, 2015.
[8] A. Azmi, "Monitoring of tool wear using measured machining forces and neuro-fuzzy modelling approaches during machining of GFRP composites," Advances in Engineering Software, vol. 82, pp. 53-64, 2015.
[9] H. Zhang, J. Zhao, F. Wang, J. Zhao, and A. Li, "Cutting forces and tool failure in high-speed milling of titanium alloy TC21 with coated carbide tools," Proceedings of the Institution of Mechanical Engineers, Part B: Journal of Engineering Manufacture, vol. 229, no. 1, pp. 20-27, 2015.
[10] X. Li, A. Djordjevich, and P. K. Venuvinod, "Current-sensor-based feed cutting force intelligent estimation and tool wear condition monitoring," IEEE Transactions on Industrial Electronics, vol. 47, no. 3, pp. 697-702, 2000.
[11] C. Madhusudana, H. Kumar, and S. Narendranath, "Face milling tool condition monitoring using sound signal," International Journal of System Assurance Engineering and Management, vol. 8, no. 2, pp. 1643-1653, 2017.
[12] J. Ratava, M. Lohtander, and J. Varis, "Tool condition monitoring in interrupted cutting with acceleration sensors," Robotics and Computer-Integrated Manufacturing, vol. 47, pp. 70-75, 2017.
[13] A. Gouarir, G. Martínez-Arellano, G. Terrazas, P. Benardos, and S. Ratchev, "In-process tool wear prediction system based on machine learning techniques and force analysis," Procedia CIRP, vol. 77, pp. 501-504, 2018.
[14] S. He, Y. Liu, J. Chen, and Y. Zi, "Wavelet transform based on inner product for fault diagnosis of rotating machinery," in Structural health monitoring: Springer, 2017, pp. 65-91.
[15] J. Liu, Y. Hu, B. Wu, and C. Jin, "A hybrid health condition monitoring method in milling operations," The International Journal of Advanced Manufacturing Technology, vol. 92, no. 5-8, pp. 2069-2080, 2017.
[16] N. E. Huang et al., "The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis," Proceedings of the Royal Society of London. Series A: mathematical, physical and engineering sciences, vol. 454, no. 1971, pp. 903-995, 1998.
[17] J. S. Richman and J. R. Moorman, "Physiological time-series analysis using approximate entropy and sample entropy," American Journal of Physiology-Heart and Circulatory Physiology, vol. 278, no. 6, pp. H2039-H2049, 2000.
[18] M. Costa, A. L. Goldberger, and C.-K. Peng, "Multiscale entropy analysis of complex physiologic time series," Physical review letters, vol. 89, no. 6, p. 068102, 2002.
[19] M. Costa, A. L. Goldberger, and C.-K. Peng, "Multiscale entropy analysis of biological signals," Physical review E, vol. 71, no. 2, p. 021906, 2005.
[20] A. Weremczuk, M. Borowiec, M. Rudzik, and R. Rusinek, "Stable and unstable milling process for nickel superalloy as observed by recurrence plots and multiscale entropy," Eksploatacja i Niezawodność, vol. 20, no. 2, 2018.
[21] C. E. Shannon, "Communication in the presence of noise," Proceedings of the IEEE, vol. 72, no. 9, pp. 1192-1201, 1984.
[22] H. Nyquist, "Certain topics in telegraph transmission theory," Transactions of the American Institute of Electrical Engineers, vol. 47, no. 2, pp. 617-644, 1928.
[23] T. IGARASHI, Y. TOKUNAGA, and K. OOKUMA, "Studies on the sound and vibration of a ball screw: sound characteristics of a ball screw," JSME International Journal. Ser. 3, Vibration, Control Engineering, Engineering for industry, vol. 31, no. 4, pp. 732-738, 1988.
[24] H. Li, L. Yang, and D. Huang, "The study of the intermittency test filtering character of Hilbert–Huang transform," Mathematics and Computers in Simulation, vol. 70, no. 1, pp. 22-32, 2005.
[25] L. Y. Lu, "Fast intrinsic mode decomposition of time series data with sawtooth transform," arXiv preprint arXiv:0710.3170, 2007.
[26] L. Y. Lu, "Fast Intrinsic Mode Decomposition of Time Series Data," 2008.
[27] Z. Peng, W. T. Peter, and F. Chu, "An improved Hilbert–Huang transform and its application in vibration signal analysis," Journal of sound and vibration, vol. 286, no. 1-2, pp. 187-205, 2005.
[28] G. Gai, "The processing of rotor startup signals based on empirical mode decomposition," Mechanical Systems and Signal Processing, vol. 20, no. 1, pp. 222-235, 2006.
[29] J. W. Cooley and J. W. Tukey, "An algorithm for the machine calculation of complex Fourier series," Mathematics of Computation, vol. 19, no. 90, pp. 297-301, 1965.
[30] H. Li, H. Zheng, and L. Tang, "Gear fault diagnosis based on order tracking and Hilbert-Huang transform," in 2009 Sixth International Conference on Fuzzy Systems and Knowledge Discovery, 2009, vol. 4: IEEE, pp. 468-472.
[31] S. M. Pincus, "Approximate entropy as a measure of system complexity," Proceedings of the National Academy of Sciences, vol. 88, no. 6, pp. 2297-2301, 1991.
[32] W. S. McCulloch and W. Pitts, "A logical calculus of the ideas immanent in nervous activity," The Bulletin of Mathematical Biophysics, vol. 5, no. 4, pp. 115-133, 1943.
[33] F. Rosenblatt, The perceptron, a perceiving and recognizing automaton Project Para. Cornell Aeronautical Laboratory, 1957.
[34] B. Widrow, Adaptive" adaline" Neuron Using Chemical" memistors.". 1960.
[35] Y.-M. Cheung, "k∗-Means: A new generalized k-means clustering algorithm," Pattern Recognition Letters, vol. 24, no. 15, pp. 2883-2893, 2003.
[36] M. Bernacki, P. Wlłodarczyk, and A. Golłda, "Principles of training multi-layer neural network using backpropagation algorithm," Neural Networks and Artificial Intelligence in Electronics, http://galaxy. agh. edu. pl/̃vlsi/AI/backp-t-en/backprop. html, Last Accessed: November, 2007.
[37] D. E. Rumelhart, G. E. Hinton, and R. J. Williams, "Learning representations by back-propagating errors," Nature, vol. 323, no. 6088, pp. 533-536, 1986.
[38] D. Rumelhart and G. Hinton, "R. J. Williams;“," Learning Internal Representations by Error Propagation,” in Parallel Distributed Processing, The MIT Press: Cambridge, Massachusetts, 1986.
[39] G. E. Hinton and R. R. Salakhutdinov, "Reducing the dimensionality of data with neural networks," Science, vol. 313, no. 5786, pp. 504-507, 2006.
[40] J. Masci, U. Meier, D. Cireşan, and J. Schmidhuber, "Stacked convolutional auto-encoders for hierarchical feature extraction," in International conference on artificial neural networks, 2011: Springer, pp. 52-59.
[41] K. Chen, J. Hu, and J. He, "A framework for automatically extracting overvoltage features based on sparse autoencoder," IEEE Transactions on Smart Grid, vol. 9, no. 2, pp. 594-604, 2016.
[42] L. Meng, S. Ding, N. Zhang, and J. Zhang, "Research of stacked denoising sparse autoencoder," Neural Computing and Applications, vol. 30, no. 7, pp. 2083-2100, 2018.
[43] Z. Zhang, "Derivation of backpropagation in convolutional neural network (cnn)," University of Tennessee, Knoxville, TN, 2016.
[44] Z. Wang and T. Oates, "Encoding time series as images for visual inspection and classification using tiled convolutional neural networks," in Workshops at the Twenty-Ninth AAAI Conference on Artificial Intelligence, 2015.
[45] E. Gadelmawla, M. Koura, T. Maksoud, I. Elewa, and H. Soliman, "Roughness parameters," Journal of Materials Processing Technology, vol. 123, no. 1, pp. 133-145, 2002.
[46] W. P. Yang and Y. Tarng, "Design optimization of cutting parameters for turning operations based on the Taguchi method," Journal of Materials Processing Technology, vol. 84, no. 1-3, pp. 122-129, 1998.
[47] M. Nalbant, H. Gökkaya, and G. Sur, "Application of Taguchi method in the optimization of cutting parameters for surface roughness in turning," Materials & Design, vol. 28, no. 4, pp. 1379-1385, 2007.
[48] M. S. H. Bhuiyan and I. A. Choudhury, "Investigation of tool wear and surface finish by analyzing vibration signals in turning Assab-705 steel," Machining Science and Technology, vol. 19, no. 2, pp. 236-261, 2015.
[49] A. Diniz, J. Liu, and D. Dornfeld, "Correlating tool life, tool wear and surface roughness by monitoring acoustic emission in finish turning," Wear, vol. 152, no. 2, pp. 395-407, 1992.
[50] G. Wang, L. Qian, and Z. Guo, "Continuous tool wear prediction based on Gaussian mixture regression model," The International Journal of Advanced Manufacturing Technology, vol. 66, no. 9-12, pp. 1921-1929, 2013.
[51] M. Hu, W. Ming, Q. An, and M. Chen, "Tool wear monitoring in milling of titanium alloy Ti–6Al–4 V under MQL conditions based on a new tool wear categorization method," The International Journal of Advanced Manufacturing Technology, vol. 104, no. 9-12, pp. 4117-4128, 2019.
[52] P. Malliavin and G. Letac, “傅立葉分析及頻譜分析”, 復漢出版社, 1991.
[53] S.Raschka and V. Mirjalili, “Python 機器學習”, 博碩文化, 2018.