本研究提出了熔融紡絲機台之加工參數異常診斷系統，以聚丙烯作為實驗材料，熔融紡絲機為實驗機台，其加工參數包括:螺桿溫度、齒輪溫度、模頭溫度、螺桿轉速、齒輪轉速、捲取轉速共六個控制因子，品質特性包括:細度、斷裂強度、斷裂伸長率、彈性能模數共四個品質。本研究第一部分以田口方法中的直交表規劃實驗，並利用實驗所得到之品質數據計算信號雜訊比、變異數分析以及信賴區間，再配合主成份分析即可得到多品質最佳加工參數。第二部分以最佳加工參數為基準，分別改變加工參數進行實驗，使品質數據偏離原本的最佳值，如此可以獲得異常樣本，接著利用多變量統計管制圖中的Hotelling’s T²管制圖對所有樣本進行分析。以最佳加工參數所得之實驗樣本作為最佳歷史數據並且求出管制界限，利用Hotelling’s T²方法算出T2值，若T2值超出管制界限即被判定為異常，找出異常品後再利用RAM(Runger, Alt and Montgomery)法對異常的T2值進行分解，計算出細度、斷裂強度、斷裂伸長率、彈性能模數四項品質的特徵值。第三部分利用RAM法求得的特徵值建構出決策樹分類器，將決策樹分類器分成單雙辨識、單因子異常、雙因子異常三個部分以提高分類的正確率，其中訓練樣本共210個，測試樣本共210個，結果顯示將三個分類器分開的單雙辨識的分類正確率為98.5%、單因子分類正確率為100%、雙因子分類正確率為92%，將其結合後單雙辨識的分類正確率為98.5%、單因子分類正確率為95.24%、雙因子分類正確率為91.84%，故證實本研究所提出的診斷方法，可以有效的找出異常樣本並建立一套熔融紡絲加工參數異常的診斷系統。

This study proposes a fault processing parameter diagnosis system for melt-spinning machine. Melt spinning machine is the experimental machine and polypropylene is the experimental material. The processing parameters include extruder temperature, gear pump temperature, die-head temperature, rotation speed of extruder, rotation speed of gear pump and coiling speed. The quality characteristics include fiber fineness, breaking strength, elongation at break and modulus of resilience. The first part of this study uses the orthogonal array of Taguchi method to design the experiment and uses the qualitative data which obtains from the experiment to calculate Signal-to-noise ratio, analysis of variance and confidence interval. The optimum process parameter level combination is determined from the qualitative data which comes from the experiment by using principal component analysis (PCA). The second part is based on the optimal process condition. We adjust the processing parameters separately to make the qualitative data deviate from the optimum parameter and get the abnormal sample. Next we use Hotelling’s T² control chart which is one of the multivariate statistical control chart to analyze all the samples. Use the best processing parameters as the best historical data to calculate UCL and then we can get T² value which comes from Hotelling’s T² method. The T² value is judged as abnormal if it exceeds the control limit. Find out the abnormal samples and use the RAM (Runger, Alt and Montgomery) method to decompose the abnormal T² value so that we can obtain the eigenvalue of each quality characteristic. In the third part, we use the eigenvalue to construct a decision tree classifier. To improve the classification accuracy rate, we classify the decision tree classifier into single-double identification, single factor abnormality and double factor abnormality. There are 210 samples for training and 210 samples for testing. The experimental results show that if we separate three classifier, the accuracy rate of the single-double identification classification is 98.5%, the accuracy rate of the single factor abnormality classification is 100% and the accuracy rate of the double factor abnormality classification is 92%. If we combine all result, we can get 98.5% accuracy rate of the single-double identification classification, 95.24% accuracy rate of the single factor abnormality classification, and 91.84% accuracy rate of the double factor abnormality classification. Therefore it can be confirmed that the proposed methods in this study can effectively identify abnormal samples and establish a fault processing parameter diagnosis system for melt spinning machine.

1.吳偉欽，“聚丙烯纖維染色性之探討”，絲織園地，第26期。
2.S. Kase and T. Matsuo, “Studies on melt spinning. I. Fundamental equation on the dynamics of melt spinning,” Journal of Applied Polymer Science, Vol. 3, pp. 2541-2554, 1965.
3.S. Kase and T. Matsuo, “Studies on melt spinning. II. Steady-state and transient solutions of fundamental equations compared with experimental results,” Journal of Applied Polymer Science, Vol. 11, pp. 251-287, 1967.
4.A. Dutta and V.M. Nadkarni, “Identifying critical process variable in poly(ethylene terephthalate) melt spinning,” Textile Research Journal, Vol. 54, No. 1, pp. 35-42, 1983.
5.K.W. Hutchenson, D.D. Edie and D.M. Riggs, “Radial temperature differences during the melt spinning of fibers,” Journal of Applied Polymer Science, Vol. 29, pp. 3621-3640, 1984.
6.V.B. Gupta and Y.C. Bhuvanesh, “Spinning speed-throughput rate relationships for polyester, Nylon, and polypropylene fibers,” Journal of Applied Polymer Science, Vol. 65, pp. 1773-1788, 1997.
7.M.K. Balki, C. Sayin and M. Sarıkaya, “Optimization of the operating parameters based on Taguchi method in an SI engine used pure gasoline, ethanol and methanol,” Fuel, Vol. 180, pp. 630-637, 2016.
8.C.F. Kuo, C.C. Huang, C.H. Ting, M.Y. Dong, W.L. Lan, “Research and development of a composite with transparent polypropene fiber Part I: a study of combining the Taguchi method with the analytic hierarchy process for masterbatch modification and toughening to enhance characteristics,” Textile Research Journal, 2017.
9.H. Esen and E. Turgut, “Optimization of operating parameters of a ground coupled heat pump system by Taguchi method,” Energy and Buildings, Vol. 107, pp. 329-334, 2015.
10.I. Asiltürk and S. Neşeli, “Multi response optimisation of CNC turning parameters via Taguchi method-based response surface analysis,” Measurement, Vol. 45, pp. 785-794, 2012.
11.Z.H. Xu, S.C. Wang, Z.W. Zhang, T.S. Chin, C.K. Sung, “Optimization of magnetizing parameters for multipole magnetic scales using the Taguchi method,” IEEE Transactions on Magnetics, Vol. 51, pp. 1-4, 2015.
12.J.T. Krishankant, M. Bector and R. Kumar, “Application of Taguchi method for optimizing turning process by the effects of machining parameters,” International Journal of Engineering and Advanced Technology, Vol. 2, pp. 263-274, 2012.
13.C.T. Su and L.I. Tong, “Multi-response robust design by principal component analysis,” Total quality management, Vol. 8, pp. 409-416, 1997.
14.J. Antony, “Multi-response optimization in industrial experiments using Taguchi's quality loss function and principal components analysis,” Quality and reliability engineering international, Vol. 16, pp. 3-8, 2000.
15.S. Moshat, S. Datta, S. Bandyopadhyay, P.Pal, “Optimization of CNC end milling process parameters using PCA-based Taguchi method,” International Journal of Engineering, Science and Technology, Vol. 2, pp. 95-102, 2010.
16.Q. Hu, P. Su, D. Yu, J. Liu, “Pattern-based wind speed prediction based on generalized principal component analysis,” IEEE Transactions on Sustainable Energy, Vol. 5, pp. 866-874, 2014.
17.L. Zhang, W. Dong, D. Zhang, G. Shi, “Two-stage image denoising by principal component analysis with local pixel grouping,” Pattern Recognition, Vol. 43, pp. 1531-1549, 2010.
18.Y. Zhang and W. Lenan, “An MR brain images classifier via principal component analysis and kernel support vector machine,” Progress In Electromagnetics Research, Vol. 130, pp. 369-388, 2012.
19.劉政原，熔融紡絲張力異常之加工參數辨識，台灣科技大學材料科學與工程學系碩士論文(2008)。
20.高振清，熔融紡絲機台之異常診斷，台灣科技大學材料科學與工程學系碩士論文(2007)。
21.R.L. Mason and J.C. Young, “Implementing multivariate statistical process control using Hotelling’s T² statistics,” Quality Progress, Vol. 34, pp. 71-73, 2001.
22.D. Boullosa, J.L. Larrabe, A. Lopez, M.A. Gomez, “Monitoring through T² Hotelling of cylinder lubrication process of marine diesel engine,” Applied Thermal Engineering, Vol. 110, pp. 32-38, 2017.
23.J. Kullaa, “Damage detection of the Z24 bridge using control charts,” Mechanical Systems and Signal Processing, Vol. 17, pp. 163-170, 2003.
24.H. Hotelling, “Multivariate quality control-illustrated by air testing of sample bombsights,” Techniques of Statistical Analysis, Mcgraw-Hill, New York, pp. 11-184, 1947.
25.Z. Wang and K.C.G. Ong, “Autoregressive coefficients based Hotelling’s T² control chart for structural health monitoring,” Computers & Structures, Vol. 86, pp. 1918-1935, 2008.
26.K. Senouci, K. Medles, K. Messal, L. Dascalescu, “Multivariate control for three variables of an industrial roll-type electrostatic separator,” IEEE Transactions on Industry Applications, Vol. 51, pp. 4752-4758, 2015.
27.游明新，多變量管制圖之比較與探討，雲林科技大學工業工程與管理學系碩士論文(2003)。
28.G.C. Runger, F.B. Alt and D.C. Montgomery, “Contributors to a multivariate statistical process control chart signal,” Communications in Statistics—Theory and Methods, Vol. 25, pp. 2203-2213, 1996.
29.許博昇，整合多變量管制圖與類神經網路以辨認多重品質特性出錯之研究，輔仁大學應用統計學研究所碩士論文(2005)。
30.H.Y. Ngan, G.K. Pang and N.Y. Yung, “Performance evaluation for motif-based patterned texture defect detection,” IEEE Transactions on Automation Science and Engineering, Vol. 7, pp. 58-72, 2010.
31.H.Y. Ngan, G.K. Pang and N.Y. Yung, “Automated fabric defect detection—a review,” Image and Vision Computing, Vol. 29, pp.442-458, 2011.
32.林昀萱，鄭仁傑，建構模糊決策樹及其在有交互作用之半導體資料之資料挖礦以提昇良率之研究，清華大學工業工程與工程管理研究所(2008)。
33.V. Sugumaran and K.I. Ramachandran, “Automatic rule learning using decision tree for fuzzy classifier in fault diagnosis of roller bearing,” Mechanical Systems and Signal Processing, Vol. 21, pp. 2237-2247, 2007.
34.M. Amarnath, V. Sugumaran and H. Kumar, “Exploiting sound signals for fault diagnosis of bearings using decision tree,” Measurement, Vol. 46, pp. 1250-1256, 2013.
35.R. Jegadeeshwaran and V. Sugumaran, “Comparative study of decision tree classifier and best first tree classifier for fault diagnosis of automobile hydraulic brake system using statistical features,” Measurement, Vol. 46, pp. 3247-3260, 2013.
36.F. Ye, Z. Zhang, K. Chakrabarty, X. Gu, “Adaptive board-level functional fault diagnosis using incremental decision trees,” IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, Vol. 35, pp. 323-336, 2016.
37.S. Singh and P. Gupta, “Comparative study ID3, CART and C4.5 decision tree algorithm: a survey,” International Journal of Advanced Information Science and Technology, Vol. 27, pp. 97-103, 2014.
38.F.J. Brims, T.M. Meniawy, L. Duffus, D. de Fonseka, A. Segal, J. Creaney, N. Maskell, R.A. Lake, N. de Klerk, A.K. Nowak, “A novel clinical prediction model for prognosis in malignant pleural mesothelioma using decision tree analysis,” Journal of Thoracic Oncology, Vol. 11, pp.573-582, 2016.
39.J. Ronowicz, M. Thommes, P. Kleinebudde, J. Krysiński, “A data mining approach to optimize pellets manufacturing process based on a decision tree algorithm,” European Journal of Pharmaceutical Sciences, Vol. 73, pp. 44-48, 2015.
40.A. Gepp and K. Kumar, “Predicting financial distress: A comparison of survival analysis and decision tree techniques,” Procedia Computer Science, Vol. 54, pp. 396-404, 2015.
41.R.K. Zimmerman, G.K. Balasubramani, M.P. Nowalk, H. Eng, L. Urbanski, M.L. Jackson, H.Q. McLean, E.A. Belongia, A.S. Monto, R.E. Malosh, M. Gaglani, L. Clipper, B. Flannery, S.R. Wisniewski, “Classification and regression tree (CART) analysis to predict influenza in primary care patients,” BMC infectious diseases, Vol. 16, pp. 503, 2016.
42.湯燦泰，熔融紡絲系統參數最佳化之研究，台灣科技大學材料科學與工程學系碩士論文(2000)。
43.李輝煌，田口方法:品質設計的原理與實務，高立圖書有限公司出版，2008年。
44.李德治，童惠玲，多變量分析，雙葉書廊出版，2009年。
45.李德治，童惠玲，林孟儒，統計學第二版，博碩文化出版，2012年。
46.謝劍斌，興軍亮，張立甯等人，20個視覺機械學習，佳魁資訊股份有限公司出版，2016年。
47.李德治，應用統計學，博碩文化出版，2017年。