本論文致力於彈性纖維混紡布染色條件最佳化與品質特性預測之研究，選用被染物分別為PET與PET混Lycra，其染料為分散性染料，及被染物為Nylon與Nylon混Lycra，其染料為酸性染料，染色方法為一浴二段式浸染法，而品質特性為胚布之色力度值。本論文旨在找出染色加工製程參數的最佳組合，以達成顧客要求之胚布顏色，並建構品質特性之預測系統。文中採用田口式實驗設計法做參數設計，針對會影響染色結果之機台工作溫度、染色時間、染液濃度、浴比值，作為實驗的控制因子，並選用直交表進行實驗，同時配合變異數分析來決定最佳製程條件、顯著因子及百分比貢獻度。實驗中，以胚布色力度值望小特性作為目標特性，再以確認實驗之計算來驗證實驗的再現性。透過實驗計畫法利用最佳條件所染出胚布之色力度值也更接近目標值，另外利用影響染色結果之顯著因子來建構基因演算法結合倒傳遞類神經網路的預測系統，將基因演算法與倒傳遞法結合，來找尋類神經網路的最佳連結鏈加權值組合，不但增加學習法則的效率，可減少網路訓練時對起始條件的依賴，亦增加學習法則的強健性，準確地預測胚布之色力度值，改進目前應用最廣泛的倒傳遞類神經網路學習法則，往往會受到初始權值選取的影響，只能求得局部最佳解而不能保證搜尋到全域最佳解之缺點。

This thesis’s objective is to optimize the dyeing process’s condition and predict its quality characteristics on elastic fiber blending fabrics by using pure PET, mixture of PET and Lycra, pure Nylon, and mixture of Nylon and Lycra as dyed fabrics with the disperse dye and acid dye as dyestuffs. We investigated the dyeing method of one-bath-two-section impregnation as well as the quality characteristic of the fabrics’ color strength values. Our specific aim is to find the optimum combination of processing parameters to achieve customers’ demands by proposing the Taguchi experimental design method. Parameters including machine-operating temperature values, dyeing periods, dye concentrations, and bath ratios are regarded as the control factors, which can affect dyeing results. We obtained the orthogonal array along with the analysis of variance approach to determine the optimal conditions, significant factors, and percent contributions. In the experiment, each fabric’s color strength values are chosen to be the smaller-the-better target characteristic, and confirmation experiments verify the reproducibility of the experimentation. The color strength values of dyed fabrics in optimal conditions are much closer to the target values. The back-propagation method, one of the most commonly used procedure, is known as the steepest descent method that relies on the system’s gradient; however, it does not guarantee to reach the system’s global optimum. The significant factors influencing the dyeing results are used to construct the prediction system of back-propagation neural network combined with genetic algorithms, which are based on the simulation of natural evolutionary processes. Therefore, we used the combination of back-propagation and genetic algorithms to improve the robustness of learning rule for artificial neural network.

[1]彭正中、周玲玉編著, “合成染料” , 台灣中華書局印行 (1982).
[2]謝光裕編譯,“歐美染色加工技術最心動向”,染化雜誌,第242期, Nov, p10-19 (2004).
[3]Wang, X., Bide, M., “Factors affecting the levelness of dyeing in reused acid dyebaths for nylon”, Textile Chemist and Colorist, v 30, n 4, Apr, p 45-52 (1998).
[4]Ibrahim, N.A., Youssef, M.A. et al., “Exhaust dyeing of polyester-based textiles using high-temperature-alkaline conditions”, Journal of Applied Polymer Science, v 89, n 13, Jul 15, p 3563-3573 (2003).
[5]Jahmeerbacus, M. I., Kistamah, N. et al., “Fuzzy control of dyebath pH in exhaust dyeing”, Coloration Technology, v 120, n 2,, p 51-55 (2004).
[6]Chen, J. C. , “The study of dyeing condition for normal pressure dyeable polyester fiber”, Journal of the China Textile Institute, v 7, n 3, May, p 188-195 (1997).
[7]John, W. M. , “Burke mills robotizes its dyehous”, Textile World, v 146, n 2, Jan, p 78-79 (1996).
[8]David, B., Victor, C.G. , “Disperse dyes for polyester-a new approach to compatibility”, JSDC, v196, p237-245 (1980).
[9]Hsu, M. J. , “A study on continuous dyeing of microencapsulated vat dye via taguchi method to improve on uniformity”, Journal of the China Textile Institute, v 7, n 3, May, p 202-206 (1997).
[10]George, P. M. , “Optimization of shot peening parameters using taguchi technique”, Journal of Materials Processing Technology, V 128, October 6, p 1-6 (2002).
[11]Hench, K. W., Al-Ghanim, A. M., “Application of a neural network methodology to the analysis of a dyeing operation”, Intelligent Engineering Systems Through Artificial Neural Networks, v 5, p 873-878 (1995).
[12]Khaw, John F.C., Lim, B.S. et al., “Optimal design of neural networks using the taguchi method”, Neurocomputing, v 7, n 3, Apr, p 225-245 (1995).
[13]Chien, W. T., Yao, C. F., “The development of the predictive model for metal machining parameters based on artificial neural network and genetic algorithm,” Proceedings of the 14th National Conference on Mechanical Engineering the Chinese Society of Mechanical Engineers, p 468-475(1997).
[14]Li, Q. S., Liu, D. K., Leung, A. Y. T., Zhang, N., Tam, C. M. and Yang, L. F., “Modelling of structural response and optimization of structural control system using neural network and genetic algorithm,” The Structural Design of Tall Buildings, v9, p 279-293(2000).
[15]Taj Jo Ko et al., ”Autonomous cutting parameter regulation using adaptive modeling and genetic algorithms ”, Precision Engineering, v 22, p 243-251(1998).
[16]汪輝雄編著, “纖維染色學”, 大學圖書出版社 (1988).
[17]張蘊如編著, “染整概論”, 高竹彩色印刷有限公司 (1988).
[18]劉熾章編著, “染色學”, 新學識文教出版中心 (1980).
[19]邱永亮、魏盛德、劉泰庠編譯, “染色化學（合訂本）”, 徐氏基金 會出版 (1977).
[20]李志雄、林瑞如,“系統化染色技術(ㄧ)聚酯織物染色專家系統”,染化雜誌,第218期, Nov, p43-48 (2002).
[21]李志雄、林瑞如,“系統化染色技術(二)耐隆織物染色專家系統”,染化雜誌,第219期, Dec, p32-37 (2002).
[22]莊才晉編譯,“聚酯織物的染整加工”,染化雜誌,第199期, Apr, p30-33 (2001).
[23]吳金雄編譯,“Polyester纖維用分散均染劑”,染化雜誌,第199期, Apr, p55-59(2001).
[24]呂世光, “尼龍染色中之pH調解劑之應用” ,染化雜誌, 第111期, Dec, p 78-90 (1993).
[25]謝光裕編譯,“聚酯纖維用分散染料的開發與工業化研究”,染化雜誌,第199期, Apr, p10-18 (2001).
[26]詹哲茂編譯,“聚酯纖維的酸性還原”,染化雜誌,第236期, May, p74-76 (2004).
[27]洪良龍, “染色加工製程參數最佳化之研究”, 國立台灣科技大學(2004).
[28]蘇朝墩編著, “品質工程”, 中華民國品質學會 (2002).
[29]李輝煌編著, “田口方法-品質設計的原理與實務”, 高立圖書有限公司 (2003).
[30]鍾清章編著, “田口式品質工程導論”, 中華民國品質協會發行 (2003).
[31]Sarin, S., “Teaching Taguchi's approach to parameter design”, Quality Progress, v 30, n 5, May, p 102-107 (1997).
[32]Ryan, T. P., “Taguchi’s approach to experimental design：some conerns ”, Quality Progress, Mar, p 34-36 (1988).
[33]葉怡成編著, “應用類神經網路”, 儒林圖書 (1997).
[34]葉怡成編著, “類神經網路模式應用與實作”, 儒林圖書 (2003).
[35]羅華強編著, “類神經網路－MATLAB的應用”, 清蔚科技 (2001).
[36]王進德、蕭大全, “類神經網路與模糊控制理論入門”, 全華圖書(2003).
[37]Michael N., “Artificial Intelligence¬－A Guide to Intelligent Systems”, Addison-Wesley(2005).
[38]王小平、曹立明, “遺傳言算法－理論、應用”,西安交通大學出版社(2002).
[39]周鵬程,“遺傳演算法原理與應用活用－Matlab”,全華圖書(2001).
[40]周鵬程, “智慧型計算概論－Matlab程式語言”, 全華圖書(2004).
[41]蘇木春、張孝德, “機器學習：類神經網路、模糊系統以及基因演算法則”, 全華圖書(2001).