工具機在切削過程中，主軸系統的刀具狀態將直接影響被加工件精度。目前切削刀具的狀態分析，為透過置於主軸軸承上加速規以收集加工振動訊號，進行狀態分析。然而此種振動分析法，在購置及分析成本上均構成應用於量產機台上的相當負擔。若能夠透過較低成本，例如電流感測，而能達到主軸刀具狀態的有效診斷，為業界實務之所需。
本研究基於K-means分群與馬可夫狀態，提出一電流訊號特徵選取程序，以分析刀具的狀態。在特徵分析上，首先蒐集與匯整不同動作下之切削負載電流值，將其轉成時域統計量與頻域特徵值後，可構成候選特徵集。在特徵選取上，採K-means進行候選特徵之狀態分群，結合信賴區間以確保所選特徵之分群的顯著性。最後在特徵分析上，以最小化之特徵組合獲得最大化之馬可夫狀態移轉正確率，以選出關鍵特徵作為判斷刀具狀態移轉之依據。
在研究成果上，藉由使用CNC加工機對鐵材進行重切削，以快速模擬刀具的磨耗過程之加速實驗。透過總數31件試驗工件的加工電流與精度比對分析可發現，基於提出程序所選取之3關鍵特徵，並配合其他1~2個輔助特徵值後，可獲得刀具狀態移轉推估平均正確率於兩種移轉中為78.92% 和 83.39%。因此，本方法應可用於以電流為基礎之刀具狀態之線上估測。

During cutting process, machining tool states directly affect work pieces precision. Currently, the tool states are analyzed by collecting vibration signals from accelerometers on spindle bearings. However, buying and analyzing cost of this kind of analysis method is a heavy burden for mass machines. In practice, what industrial needs is the spindle tool states can be effectively diagnosed while the sensing cost can be reduced by current sensing.
The study based on K-means clustering and Markov states proposes a current feature selection procedure for analyzing tool states. In feature extraction, a candidate feature set is formed time-domain and frequency domain features which are derived from collections of current data in various actions and loadings. In feature selection, the feature candidates are clustered by K-means and filtered by feature confidences to ensure clustering significance of states. Finally, the feature number is minimized by maximizing accuracy of Markov transition states to determine tool state.
In results, we proceeded tool aging experiments to simulate tool wearing by heavy-duty cutting in a CNC machine. Based on the data of working current and accuracy of 31 work pieces, the mean accuracies of the two tool state estimation are 78.92% and 83.39% according to the derived three key features and other one or two assisting features. Therefore, this current-based feature selection method can be applied to estimate tool states.

摘要.......................................................i
Abstract..................................................ii
致謝.....................................................iii
目錄......................................................iv
圖目錄....................................................vi
表目錄..................................................viii
第一章 緒論................................................1
1.1 研究背景...............................................1
1.2 研究目的...............................................2
1.3 論文架構...............................................4
第二章 理論方法...........................................5
2.1 文獻探討...............................................5
2.2提出方法................................................7
2.2.1系統架構圖............................................7
2.2.2時域分析法............................................8
2.2.3頻域分析法............................................9
2.2.4馬可夫狀態...........................................10
2.2.5 K-means演算法.......................................11
2.2.6特徵評估指標.........................................12
第三章 案例分析...........................................14
3.1 案例說明..............................................14
3.1.1 實驗設計與架構......................................14
3.1.2案例假設與限制.......................................17
3.2 實驗結果分析..........................................19
3.2.1 資料處理程序........................................19
3.2.2 資料分析步驟........................................19
3.3 特徵組合分析..........................................48
第四章 結論...............................................50
4.1分類結果...............................................50
4.2未來研究方向...........................................50
參考文獻..................................................52

[1] 2010年全球工具機市場動態分析，ITIS智網研院，IEK ITIS計畫，李妍慧、劉信宏， 2011年5月。
[2] 「CNC控制診斷應用」研討會， “CNC控制診斷應用技術”，PMC－楊信生工程師，2010年6月。
[3] V. J. Johnson, A. Moosman, and P. Cotter, “A Hierarchical Model for Estimating the Early Reliability of Complex Systems,” IEEE Trans. on Reliability, vol. 54, no. 2, pp. 224-231, 2005.
[4] D. Wang and K. S. Trivedi, “Reliability Analysis of Phased-Mission System with Independent Component Repairs,” IEEE Trans. on Reliability, vol. 56, no. 3, pp. 540-551, 2007.
[5] F.-T. Cheng, H.-C. Huang, and C.-A. Kao, “Dual-Phase Virtual Metrology Scheme,” IEEE Transactions on Semiconductor Manufacturing, vol. 20, no. 4, pp. 566-571, November 2007.
[6] 應用振動訊號與模糊演算法偵測刀具狀態之研究，國立中興大學機械工程學系所，林敬偉，2008。
[7] 應用類神經網路與振動訊號之微銑刀具狀態偵測系統開發，國立中興大學機械工程學系所，謝萬澔，2008。
[8] 利用刀具狀態監測探討鑽頭不同磨耗的振動特徵，國立高雄應用科技大學模具工程系所，賴姿吟，2008。
[9] 主軸振動與聲射訊號於微銑刀具磨耗監測之應用研究，國立中興大學機械工程學系所，黃耀賢，2009。
[10] 刀具振動監控系統之研究與建置，國立彰化師範大學電機工程學系，葉展碩，2010。
[11] 高速主軸振動監控與記錄系統之研製，國立彰化師範大學電機工程學系，廖誌祥，2010
[12] 端銑刀臨界磨耗之線上即時監控方法研究，中原大學機械工程研究所，蘇聰，2010
[13] 工具鋼SKD61鑽削製程即時監測，中興大學機械工程學系所，何春福，2008
[14] T. Doguer and J. Strackeljan, “Vibration Analysis using Time Domain Methods for the Detection of small Roller Bearing Defects,” SIRM 2009 - 8th International Conference on Vibrations in Rotating Machines, Vienna, Austria, Paper ID-16, February 23–25, 2009.
[15] H. Ahmadi and K. Mollazade, “A practical approach to electromotor fault diagnosis of Imam Khomaynei silo by vibration condition monitoring,” African Journal of Agricultural Research, vol. 4, no. 4, pp. 383-388, April 2008.
[16] B. Li, M.-Y. Chow, Y. Tipsuwan, and James C. Hung, “Neural-Network-Based Motor Rolling Bearing Fault Diagnosis,” IEEE Transactions on Semiconductor Manufacturing, vol. 47, no. 5, pp. 1060-1069, 2000.
[17] C. S. Tyagi, “A Comparative Study of SVM Classifiers and Artificial Neural Networks Application for Rolling Element Bearing Fault Diagnosis using Wavelet Transform Preprocessing,” Proceedings of World Academy of Science, Engineering and Technology, vol. 33, pp. 319-227, 2008.
[18] B. Samanta, K. R. Al-Balushi, and S. A. Al-Araimi, “Bearing Fault Detection Using Artificial Neural Networks and Genetic Algorithm,” Journal on Applied Signal Processing, pp.366–377, 2004.
[19] R. G. Silva, R. L. Reuben, K. J. Baker, and S. J. Wilcox, “Tool wear monitoring of turning operations by neural network and expert system classification of a feature set generated from multiple sensors,” Mechanical Systems and Signal Processing, vol. 12, issue 2, pp. 319-332,1998.
[20] Y. M. Niu, Y. S. Wong, and G. S. Hong, “An intelligent sensor system approach for reliable tool flank wear recognition,” The International Journal of Advanced Manufacturing Technology, Vol. 14, Number 2, pp.77-84, 1998.
[21] K. Zhu, Y. S. Wong, and G. S. Hong, “Multi-category micro-milling tool wear monitoring with continuous hidden Markov models,” Mechanical Systems and Signal Processing, vol. 23, issue 2, pp. 547-560, 2009.
[22] D. D’Addona and R. Teti, “Tool wear analysis in the composites machining through hierarchical and non-hierarchical clustering approaches,” Innovative Production Machines and System Conference, 2009.
[23] X. Yang, “Wear state recognition of drills based on K-means cluster and radial basis function neural network,” International Journal of Automation and Computing, vol. 7, no. 3, pp.271-276, 2010.
[24] 機械工業雜誌，291期，民國96年6月號，"刀具破壞之探討"，金屬工業研究發展中心，精微成形研發處，馬寧元、李新中。
[25] K. Potter, “Visualization of Statistical Uncertainty,” May 3, 2006