臺灣博碩士論文加值系統

Inconel 718 都使用在高溫環境、高速航空飛行器及超高音速飛彈武器系統上，是一種在傳統加工上不好加工的金屬，於是本研究採用田口法 L9 直交表進行線切割放電加工機參數加工分析，控制因子採用四個參數，每個控制因子採三個水準。分別探討對表面粗糙度與材料去除率單一目標品質的影響。並以田口法找出之各單一品質最佳組合參數。
應用田口法在機械加工時，大多為尋找單一目標品質之製程參數最佳化的組合。但田口法分析之因子水準在多重目標品質要求下會產生相互衝突，故本文提出結合田口法與灰關聯分析，求得多目標品質之最佳參數組合，以改善田口法之衝突。
傳統田口模式分析只提供一組較佳之加工參數組合，而無法預測非水準值實驗之結果。本文提出結合田口法實驗資料及類神經網路學習之優點，建構「階段式田口類神經網路模式」，以改善田口法之限制。
研究結果顯示，在表面粗糙度及材料去除率為双目標品質要求時，灰關聯度分析之較佳參數水準組合為 A1、B3、C1、D1，其表面粗糙度為 Ra 2.75 μm，而 L9 直交表實驗中，以 Run 1 之雙目標灰關聯度最佳，其表面粗糙度平均值為 2.80 μm，改善率為 1.8 %。同時在灰關聯分析双目標品質要求時，灰關聯度分析之較佳參數水準組合為 A1、B3、C1、D1，其材料去除率為 0.00190 g/s，而 L9 直交表實驗中，以 Run 1 之雙目標灰關聯度最佳，其材料去除率為 0.00123 g/s，改善率為 54.5 %。
經過階段式田口類神經網路模式之分析結果顯示，第一階段之初步網路，由於網路之訓練例不多，所以網路預測效果不佳；第二階段之精確網路可提供全域的較佳預測結果。

Inconel 718 is commonly used in high-temperature environments, high-speed aerospace vehicles, and hypersonic weapon systems; however, this material is difficult to process using conventional machining technology. This study adopted a Taguchi L9 orthogonal array for the analysis of parameters used in the wire electrical discharge machining of Inconel 718. The control factors included four parameters with three levels each. We explored the influence of these factors on surface roughness and material removal rate as single objective qualities in order to identify optimal parameter combinations for each of the quality characteristics.
The Taguchi method is generally used to determine optimal process parameter combinations for single objective qualities in machining, due to the fact that multi-objective quality requirements often conflict with one another. To resolve this issue, we combined the Taguchi method with Grey relational analysis to obtain optimal combinations of parameters.
Conventional Taguchi models are able to produce only one optimal parameter combination, such that they are unable to predict the results of experiments that do not use level values. This study developed a staged Taguchi neural network model, to exploit the advantages of using experimentally derived data while taking advantage of the learning capacity of neural networks to overcome the limitations of the Taguchi method.
Adopting surface roughness and material removal rate as dual objectives, the optimal parameter combination derived using Grey relational analysis was A1, B3, C1, and D1, which resulted in a mean surface roughness of Ra 2.75 μm. In the L9 orthogonal array experiments, Run 1 produced the best dual-objective grey relational grade with a mean surface roughness of 2.80 μm and an improvement rate of 1.8 %. The optimal parameter combination derived using Grey relational analysis was A1, B3, C1, and D1, which resulted in a mean material removal rate of 0.00190 g/s. In L9 orthogonal array experiments, Run 1 produced the best dual-objective Grey relational grade, with a mean material removal rate of 0.00123 g/s and an improvement rate of 54.5 %.
Analysis of the staged Taguchi neural network revealed that a lack of adequate training examples in the first stage of the preliminary network led to poor predictive power. In contrast, the accuracy of the network in the second stage provided better global prediction results.

摘要 i
Abstract iii
誌謝 v
目錄 vi
表目錄 vii
圖目錄 viii
第一章 前言 1
第二章 基本原理 8
2.1放電加工原理 8
2.2 線切割放電加工原理 9
2.3 放電加工材料去除機制 10
2.4 放電加工參數 12
2.5 田口方法 14
2.6 Inconel 718 之發展與特性 17
第三章 實驗步驟與方法 19
3.1實驗規劃： 19
3.2 實驗設備 20
3.3 實驗準備 21
3.3.1 實驗材料 21
3.3.2 實驗參數： 26
3.4 粗糙度實驗方法： 29
3.5 材料去除率實驗方法： 32
3.6實驗結果與田口分析 33
3.6.1表面粗糙度： 33
3.6.2 材料去除率 37
3.7 灰關聯分析多目標品質 39
3.7.1灰關聯分析之實驗驗證 43
第四章 線切割放電加工之多目標預測模式 45
第五章 結論 54
第六章 未來展望 56
參考文獻 57

1.K. A. Hekman , and S. Y. Liang, “Feed rate optimization and depth of cut control for productivity and part parallelism in grinding”, Mechatronics, Vol. 9, pp. 447-462,1999.
2.Y. S. Xu, J. Hu, and B. Lin, Z.C. Li, “Studies on the surface quality of the unsteady-state grinding technique”, Journal of Materials Processing Technology, Vol. 129, pp. 364-368, 2002.
3.P. L. Tso, “Study on the grinding of Inconel 718”, Journal of Materials Processing Technology, Vol. 55, pp. 421-426, 1995.
4.Z. Shi, and S. Malkin “An investigation of grinding with electroplated CBN wheels”, Annals of CIRP, Vol. 52, pp. 267-271, 2003.
5.Bi Zhang, X. L. Zheng, H. Tokura, and M. Yoshikawa, “Grinding induced damage inceramics”, Journal of Materials Processing Technology, Vol. 132, pp. 353-364, 2003.
6.X. P. Xu, Y. Li, W. Y. Zeng, and L. B. Li, “Quantitative analysis of the loads acting on the abrasive grits in the diamond sawing of granites”, Journal of Materials Processing Technology, Vol. 129, pp. 50-55, 2002.
7.S. L. Chen, B. H. Yan, and F. Y. Huang, “Influence of kerosene and distilled water asdielectrico on the electric discharge machining characteristics of Ti-6Al-V”, Journal of Materials Processing Technology, Vol. 87, pp. 48-55, 1999.
8.J. C. Rebelo, A. M. Dias , and J. L. Lebrun, “Influence of EDM pulse energy on the surface integrity of matensitic steels”, Journal of Materials Processing Technology, Vol. 84, pp. 90-96, 1998.
9.S. H. Lee, and X. P. Li, “Study of the effect of machining parameters on the machining characteristics in electrical discharge machining of tungsten carbide”, Journal of Materials Processing Technology, Vol. 115, pp. 344-358, 2001.
10.J. L. Lin, and C. L. Lin, “The use of the orthogonal array with grey relational analysis to optimize the electrical discharge machining process with multiple performance characteristics”, International Journal of Machine Tools and Manufacture, Vol. 42, pp. 237-244, 2002.
11.B. H. Yan, F. Y. Huang, H. M. Chow, and J. Y. Tsai, “Micro-hole machining of carbide by electric discharge machining”, Journal of Materials Processing Technology, pp. 139-145, 1999.
12.E. Uhlmann, S. Piltz, and S. Jerzembeck, “Micro-machining of cylindrical parts by electrical discharge grinding”, Journal of Materials Processing Technology, Vol. 160, pp. 15-23, 2005.
13.于劍平，“放電加工表面特性之研究＂，碩士論文，國立成功大學機械工程系，2002年。
14.林耀乾，“鎢銅複合材料放電加工特性與多重品質之研究＂，碩士論文，國防大學中正理工學院兵器系統工程所，2002年。
15.蘇建嘉，“小徑深孔放電加工特性之研究＂，碩士論文，國立台灣大學機械工程所，2003年。
16.黃繼輝，“應用田口法於五軸CNC銑削加工製程之最佳化參數設計”，國立台灣科技大學機械工程技術研究所碩士論文，2008年。
17.楊國榮，“以田口法探討導線架電漿改製程之最佳化”，華梵大學機電工程學系論文，2005年。
18.Yang, C. B. and Chiang, H. L., “Using Gray-based Taguchi Method to Construct Multi-Objective Optimal Model in Super-Resolution Near-Field Photolithography”, Scanning, Vol. 35(1), pp. 47-58, 2013.
19.M. T. Haganm, H. B. Demuth, and M. H. Beale, “Neural networks Design”, Thomson Learning, 1996.
20.W. T. Miller, R. S. Sutton, and P. J. Werbos, “Neural networks for control”, MIT Press, 1990.
21.D.A.White and D.A.Sofge Edited, “Hanbook of intelligent control”, Neural, Fuzzy, and Adaptive Approaches, Van Nostrand Reinhold, 1992.
22.W.S. Lin, B. Y. Lee, and C. L. Wu, “Modeling the surface Roughness and Cutting Force for Turning”, Elaevier Science, 1999.
23.吳俊中，“類神經模糊系統與遺傳演算法於多目標加工參數最佳化之應用”，國立海洋大學機械工程研究所碩士論文，2000年。
24.張渭川編譯，“放電加工的結構與實用技術”，全華圖書，1986年。
25.鳳誠三郎、蒼藤尚雄，“放電加工”，復漢出版社，1984年。
26.蘇品書編著，“線切割放電加工”，復漢出版社，2000年。
27.黃士瑋，“使用電泳沉積研磨加工改善放電微孔精度之研究”，碩士論文，中央大學機械工程研究所，2003年。
28.賴耿陽，“機械加工計測技術”，初版，復文書局，1987年。
29.M. Alauddin, M. A. Mazid, M. A. El Baradi, and M. S. J. Hashmi, “Cutting forces in the end milling of Inconel 718”, Journal of Materials Processing Technology, Vol. 77, pp. 153-159, 1998.
30.M. Alauddin, M. A. El Baradie, and M. S. J. Hashmi, “Modelling of utting force in end milling Inconel 718”, Journal of Materials Processing Technology, Vol. 58, pp. 100-108, 1996.
31.楊慶宜，“MAR-M247 超合金高速銑削性探討”，國立中正大學機械工程所，1998年。
32.國家中山科學研究院材料手冊。
33.李輝煌，“田口方法-品質設計的原理與實務”，第四版，高立圖書有限公司，2011年。
34.Wen, K. L., Huang, T. L., Chang, W. C., Chang, T. C., Lai, C. J., “Grey relational model method and application”, Gau Lih Book Co., LTD., Taipei, 2003.
35.溫坤禮、黃宜豊、張偉哲、張廷政、游美利、賴家瑞，“灰關聯模型方法與應用”，高立圖書有限公司，2003年。
36.Z. C. Lin and D. Y. Chang, “Application of a neural network machine learning model in the selection system of sheet bending tooling”, International Journal of Artificial Intelligence in Engineering, Vol.10, pp.21-37, 1996.