(3.238.249.17) 您好!臺灣時間:2021/04/14 13:13
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果

詳目顯示:::

我願授權國圖
: 
twitterline
研究生:林家平
論文名稱:以接續式擬均勻設計進行產品與程序最適條件之開發
論文名稱(外文):Product and process development via sequential pseudo-uniform design
指導教授:張志雄張志雄引用關係
學位類別:碩士
校院名稱:大同大學
系所名稱:化學工程研究所
學門:工程學門
學類:化學工程學類
論文種類:學術論文
論文出版年:2003
畢業學年度:91
語文別:中文
論文頁數:85
中文關鍵詞:均勻設計
相關次數:
  • 被引用被引用:0
  • 點閱點閱:67
  • 評分評分:系統版面圖檔系統版面圖檔系統版面圖檔系統版面圖檔系統版面圖檔
  • 下載下載:0
  • 收藏至我的研究室書目清單書目收藏:0
應用傳統實驗設計方法來尋找產品或程序之最適操作條件,因為其所架構之模式通常為一階或二階模式,因而控制因子的測試範圍將較會受到限制。若控制因子之測試範圍較廣且程序為非線性,如何有效進行實驗設計以建立此系統的代表性模式並求取此系統之最適操作條件是本研究的主要目標。均勻設計能夠處理 個因子且每一因子可分為 水平總共只需進行 個實驗的設計。如果每一個實驗的費用是昂貴的,通常先採用較少分割水平的均勻設計來進行實驗。若上述的實驗數據無法架構一個足以代表系統的類神經網路模式時,表示實驗數據資訊尚不夠充足而需要更多的實驗數據來架構出一正確的系統模式。我們提出所謂的接續式均勻實驗設計來安排如何接續先前之均勻設計所需之新的實驗佈點以提供更充足的系統資訊。經由充足的實驗數據並藉由隨機搜尋法、模糊分類與資訊自由能分析來確定程序之最佳操作條件。為了驗證所提出之實驗設計方法的有效性,我們以修飾的Himmelblau函數的最小值問題、盤尼西寧發酵程序和批式系列反應程序求取最適溫度操作路徑問題進行相關驗證,並獲得滿意的結果。

Application of classical experiment design methods to find optimal operating conditions for a process, the testing regions of the control factors are usually located in limited spans. However, if the testing regions are broad and the process is nonlinear, how to design the experiments efficiently and build a representative model based on the experiment data is the main objective of this work. The uniform design method can design a total of q experiments for a process with n factors and q levels in each factor. If the cost of each experiment is high, low partitioned levels are usually proposed first to do the experiments. However, if a reliable artificial neural network (ANN) model cannot be built based on the obtained experimental data, we developed a sequential pseudo-uniform design method to locate additional experiments in the experimental region. Furthermore, information analysis was used to confirm the identified ANN model is consistent with the process. Applications of the proposed method to a benchmark problem, namely minimization of the modified Himmelblau function, a penicillin fermentation process and a series reaction in batch reactor to obtain an optimal operating trajectory were examined. The results prove that the product and process development based on the proposed method requires a reasonable number of experiments.

CONTENTS
ACKNOWLEDGMENTS i
ABSTRACT (in English) ii
ABSTRACT (in Chinese) iii
CONTENTS iv
LIST OF FIGURES vii
LIST OF TABLES xii
NOTATION xiii
CHAPTER 1 INTRODUCTION 1
1.1 Statistical Experimental Design 1
1.2 Literature 3
1.3 Motive 5
1.4 Organization 6
CHAPTER 2 THEORY AND RESEARCH METHOD OF PRODUCT OR PROCESS DEVELOPMENT OPTIMUM EXPERIMENTAL DESIGN 8
2.1 Experimental Design 8
2.1.1 Full Factorial Experimental Design 9
2.1.2 Orthogonal Experimental Design 11
2.1.3 Uniform Design 16
2.1.4 Sequential Pseudo-Uniform Design 18
2.2 Artificial Neural Network 22
2.2.1 Feedforward artificial neural network structure 22
2.2.2 Feedforward artificial neural network algorithm 25
2.3 Optimization Systems 27
2.3.1 Random Search Method 28
2.3.2 Illustration 30
2.4 Information Index 33
2.4.1 Fuzzy Classification 33
2.4.2 Information entropy 34
2.4.3 Information Enthalpy 36
2.4.4 Information Free Energy 36
2.4.5 Illustration 37
2.5 Products or Process Optimization 38
CHAPTER 3 CASE STUDIES 44
3.1 Case 1 44
3.2 Case 2 52
3.3 Case 3 57
3.4 Case 4 62
3.5 Case 5 64
3.6 Case 6 70
3.7 Case 7 75
CHAPTER 4 CONCLUSION 80
REFERENCES 83
LIST OF FIGURES
Figure 2.1 3 full factorial experimental design 11
Figure 2.2 Uniform design and sequential pseudo-uniform design(○: uniform design points; △: sequential pseudo-uniform design points) 20
Figure 2.3 Feedforward artificial neural networks 23
Figure 2.4 The multilayer feedforward artificial neural network architecture 24
Figure 2.5 Flow chart for the random search method 31
Figure 2.6 Random search algorithm 32
Figure 2.7 (a) Three-dimensional mesh; (b) group data points against the contour background 38
Figure 2.8 (a) Information entropy vs. cluster; (b) entropy multiplied by temperature vs. cluster; (c) enthalpy vs. cluster; (d) free energy vs. cluster 40
Figure 2.9 Flow chart for sequential pseudo-uniform design42
Figure 3.1 Three-dimensional modified Himmelblau function45
Figure 3.2 The contour of modified Himmelblau function 46
Figure 3.3 (a) The first experiments of uniform design (b) The first new inserting experiment (testing points of 1st Run ●; 2nd Run ○; 3rd Run ▼; 4th Run ▽) 48
Figure 3.4 Optimal experimental design of modified Himmelblau function. ((a) experimental arrangement points: training ○ and testing ▲ (b) testing point error (c) location of the best model input value (d) objective value of the best model input value (e) information index) 49
Figure 3.5 Optimal experimental design of modified Himmelblau function following Figure 3.4. ((a) experimental arrangement points: training ○ and testing ▲ (b) testing point error (c) location of the best model input value (d) objective value of the best model input value (e) information index) 51
Figure 3.6 Figure 3.6 Uniform points near optimal model input point of modified Himmeblau function 53
Figure 3.7 Optimal experimental design of modified Himmelblau function with noise. ((a) experimental arrangement points: training ○ and testing ▲ (b) testing point error (c) location of the best model input value (d) objective value of the best model input value (e) information index) 54
Figure 3.8 Optimal experimental design of modified Himmelblau function with noise following Figure 3.7. ((a) experimental arrangement points: training ○ and testing ▲ (b) testing point error (c) location of the best model input value (d) objective value of the best model input value (e) information index) 55
Figure 3.9 The temperature trajectories of three temperature collocation points of the batch penicillin synthesis reaction59
Figure 3.10 Optimal experimental design of the batch penicillin synthesis reaction. ((a) experimental arrangement points: training ○ and testing ▲ (b) testing point error (c) location of the best model input value (d) objective value of the best model input value (e) information index) 60
Figure 3.11 The optimal temperature trajectory of three temperature collocation points of the batch penicillin synthesis reaction 61
Figure 3.12 Uniform points near optimal model input of the batch penicillin synthesis reaction 63
Figure 3.13 Optimal experimental design of the batch penicillin synthesis reaction with noise. ((a) experimental arrangement points: training ○ and testing ▲ (b) testing point error (c) location of the best model input value (d) objective value of the best model input value (e) information index) 65
Figure 3.14 Optimal experimental design of the batch penicillin synthesis reaction with noise following Figure 3.13. ((a) experimental arrangement points: training ○ and testing ▲ (b) testing point error (c) location of the best model input value (d) objective value of the best model input value (e) information index) 66
Figure 3.15 The optimal temperature trajectory of three temperature collocation points of the batch penicillin synthesis reaction with noise 67
Figure 3.16 The optimal temperature trajectory of the four temperature collocation points of the batch penicillin synthesis reaction 69
Figure 3.17 Optimal experimental design of the batch series reaction (the 1st point of the temperature collocation is set at 125℃). ((a) experimental arrangement points: training ○ and testing ▲ (b) testing point error (c) location of the best model input value (d) objective value of the best model input value (e) information index) 72
Figure 3.18 Optimal experimental design of the batch series reaction following Figure 3.17 (the 1st point of the temperature collocation is set at 125℃). ((a) experimental arrangement points: training ○ and testing ▲ (b) testing point error (c) location of the best model input value (d) objective value of the best model input value (e) information index) 73
Figure 3.19 The optimal temperature trajectory of the 3 temperature collocations of the batch series reaction (the first point is set at 125℃) 74
Figure 3.20 Optimal experimental design of the batch series reaction. ((a) experimental arrangement points: training ○ and testing ▲ (b) testing point error (c) location of the best model input value (d) objective value of the best model input value (e) information index) 76
Figure 3.21 Optimal experimental design of the batch series reaction following Figure 3.20. ((a) experimental arrangement points: training ○ and testing ▲ (b) testing point error (c) location of the best model input value (d) objective value of the best model input value (e) information index) 77
Figure 3.22 Optimal temperature trajectory of three temperature collocations of the batch series reaction 78
LIST OF TABLES
Table 2.1 L Orthogonal Table 13
Table 2.2 L of Conversion Experiment 15
Table 2.3 Uniform Table 17
Table 2.4 U of Production of Chemical Industry 18
Table 3.1 Relationship Between The Run and Node of Case 1 51
Table 3.2 Relationship Between The Run and Node of Case 2 56
Table 3.3 Relationship Between The Run and Node of Case 3 62
Table 3.4 Relationship Between The Run and Node of Case 4 64
Table 3.5 Relationship Between The Run and Node of Case 5 68
Table 3.6 Parameters and Given Conditions for The Batch Series Reaction 71
Table 3.7 Relationship Between The Run and Node of Case 6 75
Table 3.8 Relationship Between The Run and Node of Case 7 76

References
1. Bezdek, J. C.; Ehrlich, R.; Full, W. FCM: The Fuzzy c-Means Clustering Algorithm. Comput. Geosci. 1984, 10, 191.
2. Box, G.; Draper, N. R. Empirical Model-Building and Response Surface; Wiley: New York, 1987.
3. Chen, J.; Sheui, R. G. Using Taguchi’s Method and Orthogonal Function Approximation to Design Optimal Manipulated Trajectory in Batch Process. Ind. Eng. Chem. Res. 2002, 41, 2226.
4. Cheng, B.; Zhu, N.; Fan, R.; Zhou, C.; Zhang, G.; Li, W.; Ji, K. Computer Aided Optimum Design of Rubber Recipe using Uniform Design. Polymer Testing 2002, 21, 83.
5. Chen, J.; Wong, D. S. H.; Jang, S. S. Product and Process Development using Artificial Neural- Network Model and Information Analysis. AIChE Journal 1998, 44, 876.
6. Chu, J. Z.; Shieh, S. S.; Jang, S. S.; Chien, C. I.; Wan, H. P.; Ko, H. H. Constrained Optimization of Combustion in a Simulated Coal-Fired Boiler using Artifical Neural Network Model and Information. Fuel 2003, 82, 693.
7. Fang, K. T. Uniform Design: Application of Number-Theoretic Methods in Experimental Design. Acta Math. Appl. Sin. 1980, 3, 363.
8. Fang, K. T.; Hong, Q. A Note on Construction of Nearly Uniform Designs with Large Number of Runs. Statistics & Probability Letters 2003, 61, 215.
9. Fukunaga, K. Introduction to Statistical Pattern Recognition; Academic Press: Boston, 1990.
10. Hagan, M. T.; Demuth, H. B. Neural Network Design; Pws Publishing: Boston, 1995.
11. Jang, J. S. R.; Sun, C. T.; Mizutani, E. Neural-Fuzzy and Soft Computing; Prentice-Hall International: USA, 1997.
12. Khaw, J. F. C.; Lim, B. S.; Lim, L. E. N. Optimal Design of Neural Networks using the Taguchi Method. Neurocomputing 1995, 7, 225.
13. Kumar, A.; Motwani, J.; Otero, L. An Application of Taguchi’s Robust Experimental Design Technique to Improve Service Performance. The International Journal of Quality & Reliability Management 1996, 85.
14. Liang, Y. Z.; Fang, K. T.; Xu, Q. S. Uniform Design and its Applications in Chemistry and Chemical Engineering. Chemometrics and Intelligent Laboratory Systems 2001, 58, 43.
15. Lin, Z.; Liang, Y. Z.; Jiang, J. H.; Yu, R. Q.; Fang, K. T. Uniform Design Applied to Nonlinear Multivariate Calibration by ANN. Analytica Chimica Acta 1998, 370, 65.
16. Lochner, R. H.; Matat, J. E. Design for Quality: An Introduction to the Best of Taguchi and Western Methods of Statistical Experimental Design; ASQC Quality Press: Milwaukee, WI, 1990.
17. Norio, B.; Yoshio, M.; Motokaze, K.; Yasuhiro, S.; Yuttaka, Y. A Hybrid Algorithm for Finding the Global Minimum of Error Function of Neural Networks and its Applications. Neural Networks 1994, 7, 1253.
18. Roy, R. K. A Primer on the Taguchi Method; Van Nostrand & Reinhold: New York, 1990.
19. Shannon, C. E. A Mathematical Theory of Communication. Bell Syst. Tech. J. 1948, 27, 379.
20. Shannon, C. E.; Weave, W. The Mathematical Theory of Communication; Univ. of Illinois Press: Urbana, 1949.
21. Solis, F. J.; Wets, J. B. Minimization by random search technique. Mathematics of Operations Research 1981, 6, 19.
22. Taguchi, G. Introduction to Quality Engineering; Asian Productivity Organization: Japan, 1986.
23. 方開泰與馬長興,正交與均勻試驗設計;香港浸會大學:香港,2000

QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top
系統版面圖檔 系統版面圖檔