工具機在工業的生產上是不可或缺的，而齒輪箱是工具機常用的零組件，若是齒輪箱故障可能會導致生產線不可預知的損失，因此過去幾年來已經有許多業界與學界的研究人員投入齒輪箱錯誤診斷的研究。一般而言，要建立錯誤診斷系統的數據資料庫有三個基本步驟：擷取振動數據，特徵抽取和錯誤狀態分類。在本論文中，討論四種多尺度分析算法，包括組合式多尺度熵（CMSE），組合式多尺度排列熵（CMPE），多頻帶頻譜熵（MBSE），和多尺度奇異值分解熵（MSVDE），用來抽取不同轉速以及不同錯誤的齒輪箱振動訊號的特徵。而支持向量機（SVM）和類神經網絡（NN）是作為分類器來區分齒輪箱的錯誤類型。本論文的實驗平台是由工業技術研究院（ITRI）提供。
實驗四種不同的條件，包括正常，不平衡，齒輪磨損和齒輪斷裂，收集不同的齒輪轉速從446 rpm開始，並以12 rpm為區間，一直到2121 rpm的齒輪振動訊號。
為了評估演算法在多轉速齒輪箱錯誤診斷的可行性，對振動數據進行分組辨識，每一組共取五個不同的齒輪轉速，本論文設計兩個實驗：（1）五個不同的轉速數據同時用來訓練分類器; （2）只用一個特定轉速的數據來訓練分類器。實驗結果表明，如果訓練數據來自各個不同的速度，所提出的錯誤診斷演算法預測精度非常的高（高達99.8％）。然而，如果訓練數據僅來自一個轉速，錯誤診斷演算法的預測準確率會大大降低。本論文希望這項研究可以提供一些貢獻在開發一個多轉速齒輪箱的錯誤診斷系統。

Tool machines are essential in many manufacturing processes. Gearboxes are the common used component in a tool machine. Gearbox failures could lead to unpredictable productivity losses for production facilities. Therefore, gearbox fault diagnosis has attracted significant attention from the research and engineering communities over the past decades. In general, a data-driven fault diagnosis system consists of three general steps: vibration data acquisition, feature extraction and fault condition classification. In this paper, four multiscale scale analysis algorithms including composite multiscale entropy (CMSE), composite multiscale permutation entropy (CMPE), multiband spectrum entropy (MBSE), and multiscale singular value decomposition entropy (MSVDE) are applied to extract the features of vibration signals collected from different gearbox faults. Support vector machine (SVM) and artificial neural network (NN) are used as classifiers to distinguish the fault types of gearbox respectively.
The experimental platform is provided by Industrial Technology Research Institute (ITRI). Four different conditions including normal, imbalance, tooth-wear and tooth-broken are considered in these experiments. The vibration signals of gearbox were collected for several different motor speeds from 446rpm to 2121 rpm with a resolution of 12rpm.
To evaluate the feasibility of the proposed algorithm for multi-speeds gearbox fault diagnosis. Vibration data for five different speeds were grouped and considered as the same class. Two experiments are performed in this study: (1) data used to train a classifier come from all five different speeds; (2) data used to train a classifier came from only one specified speeds. Simulation results indicate that if the training data come from all different speeds, the accuracy of prediction of the proposed diagnosis algorithm is very high (up to 99.8%). However, the accuracy of prediction of the proposed diagnosis algorithm will decrease dramatically if the training data come from only one speed. We wish this study can provide some contribution in developing a multi-speeds gearbox fault diagnosis system.

[1] 江素雲，「2013年全球工具機市場面觀」，IT IS產業評析，2013 / 07。
[2] 傅繼盈，蔣秀珍，「機械學基礎」，哈爾濱工業大學出版社，2003 / 8。
[3] Z. Xu, J. Xuan, T. Shi, B. Wu, Y. Hu“A Novel Fault Diagnosis Method of Bearing Based on Improved Fuzzy ARTMAP and Modified Distance Discriminant Technique”Expert Systems with Applications, vol. 36, pp.11801-11807, 2009.
[4] S.Wadhwani, A. K. Wadhwani, S. P. Gupta and V. Kumar, “ Detection of Bearing Failure in Rotating Machine Using Adaptive Neuro-Fuzzy Inference System,” in Power Electronics, Drives and Energy Systems, 2006. PEDES ’06. International Conference on, New Delhi, 2006.
[5] R. Yan and R.X. Gao, ”Approximate Entropy as a Diagnosis Tool for Machine Health Monitoring,” Mechanical Systems and Signal Processing, vol. 21, pp. 824-839, 2007.
[6] J. Huang, H. Pen and S. Bi, “Bispectrum Entropy Feature Extraction and Its Application for Fault Diagnosis of Gearbox,“ in Fussy Systems (FUZZ), 2010 IEEE International Conference on, Barcelona, 2010.
[7] D. Yu, Y. Yang and J. Cheng, “Application of Time Frequency Entropy Method Based on Hilbert Huang Transform to Gear Fault Diagnosis,” Measurement, vol. 40, pp. 823-830, 2007.
[8] R. Yan and R.X. Gao, ”Complexity as a Measure for Machine Health Evalution,” IEEE Transaction on, vol. 53, no. 4, pp. 1327-1334, 2004.
[9] H. Hong and M. Liang, “Fault Severity Assessment for Rolling Element Bearing Using the Lempel-Ziv Complexity and Continuous Wavelet Transform,” Journal of Sound and Vibration, vol. 320, pp. 452-468, 2009.
[10] 林庭毅、吳順德、林倪鋒,「基於多尺度方均根植與前饋式倒傳遞網路隻機械錯誤診斷系統,」於2012中華民國系統科學與工程延討論會, 基隆市, 2012.
[11] R. Hao, Z. Peng, Z. Feng and F. Chu, “Application of Support Vector Machine Based on Pattern Spectrum Entropy in Fault Diagnostics of Rolling Element Bearings,” Measurement Science and Technology, vol. 22, p. 045708, 2011.
[12] J. S. Richman and J. R. Moorman, “Physiological time-series analysis using approximate entropy and sample entropy,” Am J Physiology Heart Circulatory Physiology, Vol. 278, pp. H2039-H2049, 2000.
[13] B. Li, P. Zhang, S. Liang and G. Ren, “Feature Extraction and Selection for Fault Diagnosis of Gear Using Wavelet Entropy and Mutual Information,” in International Conference on Signal Processing, Beijin, China, 2008.
[14] L. Zong, G. Xiong, H. Liu, H. Zou and W. Guo, “Bearing Fault Diagnosis Using Multi-scale Entropy and Adaptive Neuro-Fuzzy Inference,” Expert System with Applications, vol. 37, pp. 6077-6085, 2010.
[15] J. L. Lin, J. Y. Liu, C. W. Li, L. F. Tsai and H. Y. Chung, “Motor Shaft Misalignment Detection Using Multiscale Entropy with Wavelet Denoising,” Expert System with Applications, vol. 37, pp. 7200-7204, 2010.
[16] 吳求文、王俊傑、吳順德,「基於多尺度熵、區別指標與支持向量機之旋轉機械異常診斷系統,」於 第二十八屆中國機械工程全國學術研討會,台中市,2011.
[17] 李威諭、吳求文、吳順德,「基於多尺度排序熵與k-means 分類器之錯誤診斷系統,」於 2013中華民國系統科學與工程研討會,基隆市,2012.
[18] S.D. Wu , P.H. Wu , C.W. Wu , J.J. Ding, C. C. Wang ,“Bearing Fault Diagnosis Based on Multiscale Permutation Entropy and Support Vector Machine ,” The 3nd international Conference on Mechanic Automation and Control Engineering (MACE2012), Inner Mongolia, China, 2012
[19] Y. N. Pen, J. Chen and X. L. Li, “Spectral Entropy: A Complementary Index for Rolling Element Bearing Performance degradation Assessment,” Proc. IMechE Part C: J. Mechanical Engineering Science, no. 223, pp. 1223-1231, 2009.
[20] J. Hu, R. Shao and Z. Zeng, “Method of Gear's Fault Diagnosis based on Spectral Entropy,” Journal of Mechanical Transmission, May 2007.
[21] 王俊傑、吳求文、吳順德、李易宗、吳豐泰,「多尺度頻譜熵在軸承異常監控與診斷之應用,」於第二十八屆中國機械工程學術研討會, 台中市, 2011.
[22] 吳求文,“旋轉機械線上監控與異常辨識系統,”國立臺灣師範大學機電科技學系,碩士論文,2012 6月.
[23] S. D. Wu, C. W. Wu, S. G. Lin , C. C. Wang and K. Y. Lee “Time Series Analysis Using Composite Multiscale Entropy,” in Entropy, 15, 1069-1084; doi:10.3390/e15031069,2013
[24] 吳求文、王仁浩、吳順德,「軸承錯誤診斷基於多尺度奇異值分解熵,」於 第三十屆中國機械工程學會全國學術研討會, 宜蘭縣 2013.
[25] W. C. Kao, C. K. Yu, C. P. Shen, and W. H. Chen “Electrocardiogram Analysis with Adaptive Feature Selection and Support Vector Machines,” in n Proc. IEEE Asia Pacific Conference on Circuits and Systems, Singapore 2006.
[26] W. C. Kao, M. C. Hsu and Y. Y. Yang “Local Contrast Enhancement and Adaptive Feature Extraction for Illumination-invariant Face Recognition,” Pattern Recognition, vol. 43, no. 6, pp.1736-1747, 2010.
[27] N. Saravanan, V. N. S. Kumar Siddabattuni and K. I. Ramachandran “Fault Diagnosis of Spur Bevel Gear Box Using Artificial Neural Network (ANN), and Proximal Support Vector Machine (PSVM),” Applied Soft Computing, vol. 10, pp. 344-360, 2010.
[28] B. Sreejith, A. K. Verma and A. Srividya, "Fault Diagnosis of Rolling Element Bearing Using Tme-Domain Features and Neural Networks," IEEE Proceedings of ICIIS, pp. 1-6, 2008.
[29] W. X. Sun, J. Chen, J. Q. Li “Decision Tree and PCA-based Fault Diagnosis of Rotating Machinery,” Mechanical Systems and Signal Processing, vol. 21, pp. 1300-1317, 2007.
[30] C. W. Hsu and C. J. Lin “A Comparison of Methods for Multiclass Support Vector Machines” IEEE Transactions on Neural Networks, vol. 13, no. 2, pp. 414-9425, 2002.
[31] 李易宗,“ 基於強化型Morlet轉換、解調變頻譜、多尺度熵、多頻帶頻譜熵與決策樹之齒輪箱異常診斷系統,”國立臺灣師範大學機電科技學系,碩士論文,2012 7月
[32] http://spl.mt.ntnu.edu.tw/ITRI_data_menu.html
[33] C. E. Shannon, “A Mathematical Theory of Communication,” Bell System Technical Journal, vol. 27, pp. 379-423, 623-656, Jul, Oct 1948.
[34] C. Bandt and B. Pompe, “Permutation Entropy: A Natural Complexity Measure for Time Series,” Physical Review Letter, vol. 88, pp. 174102-1-174102-4, April 2002.
[35] Y. B. Liu, Q. Long, Z. H. Feng and W. L. Liu, “Detection Method for Nonlinear and Non-stationary Signals,” Journal of Vibration and Shock, Dec 2007.
[36] R. Q Yana, Y. B. Liu, R. X. Gao “Permutation entropy : A nonlinear statistical measure for status characterization of rotary machines,” Mechanical Systems and Signal Processing, Dec 2011.
[37] G. E. Powell and I. C. Percival “A spectral entropy method for distinguishing regular and irregular motion of Hamiltonian systems,” Journal of Physics A: Mathematical and General, vol. 12, pp. 2053-2071, Nov 1979.
[38] O. Alter, P. O. Brown, and D. Botstein, "Singular value decomposition for genome-wide expression data processing and modeling," Proceedings of the National Academy of Sciences of the United States of America, vol. 97, pp. 10101-10106, Aug 29 2000.
[39] M. Costa, A. L. Goldberger and C. K. Peng, “Multiscale Entropy Analysis of Complex Physiologic Time Series,” Physical Review Letters, Vol. 89, No.6, pp. 068102-1 - 068102-4, 2002.
[40] M. Costa, A. L. Goldberger and C. K. Peng, “Multiscale Entropy Analysis of Biological Signals,” Physical Review E, Vol.71, pp. 021906-1 - 021906-18, 2005.
[41] W. Aziz , M. Arif, “Multiscale Permutation Entropy of Physiological Time Series,” on 9th International Multitopic Conference, IEEE INMIC 2005, Karachi, 2005.
[42] C.C. Chang and C.J. Lin, LIBSVM : a library for support vector machines, 2001. Software available at http://www.csie.ntu.edu.tw/~cjlin/libsvm
[43] 張斐章，張麗秋，「類神經網路」，台灣東華書局股份有限公司，2005 / 9。
[44] D. E. Rumelhart, G. E. Hinton and R. J. Williams, "Learning Internal Representation by Error Propagation," in Parallel Distributed Processing, D. E. Rumelhart and McClelland(Eds), MIT, Press, Cambridge, MA, vol. 1, pp. 318-362, 1986.
[45] 林庭毅,“ 基於多尺度方均根、多尺度熵、費雪法與倒傳遞網路之軸承錯誤診斷系統,”國立臺灣師範大學機電科技學系,碩士論文,2012 6月
[46] 趙景明、梁淑芳，「導入LM法之平行倒傳遞演算法」，資訊管理展望，第8卷，第2期，第85~108頁，2006年12月。
[47] M. T. Hagan and M. B. Menhaj, "Training Feedforward Networks with the Marquardt Algorithm," IEEE Trans. Neural Networks, vol. 5, pp. 989-993, 1994