(18.204.227.34) 您好!臺灣時間:2021/05/14 09:48
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:劉坤志
研究生(外文):Liu Kun Chih
論文名稱:線上及時批次製程監控
論文名稱(外文):On-line Batch Process Monitoring
指導教授:陳榮輝陳榮輝引用關係
指導教授(外文):Junghui Chen
學位類別:碩士
校院名稱:中原大學
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:2001
畢業學年度:89
語文別:中文
論文頁數:73
中文關鍵詞:統計制程管制主成份分析批次制程
外文關鍵詞:PCASPCBatch ProcessPLS
相關次數:
  • 被引用被引用:1
  • 點閱點閱:145
  • 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:1

目前的批次製程MSPC研究中,僅有MacGregor(1995)所提出的MPCA(Multi-way Principal Component Analysis)和MPLS(Multi-way Partial Least Squares)兩種方法。MPCA主要原理是將批次製程的三維立方體數據拆成二維的平面數據,再帶入傳統的PCA中做運算,這樣便可以將多變數間主要分佈的方向抽取出來,以達到壓縮變數,萃取變數的目的。MPLS如同MPCA一樣,是將三維立方體數據拆成二維的平面數據,再帶入PLS中做運算。但它強調操作變數與產品品質之間的關係,希望能在兩者的分佈方向上找到一個最適值。但是此兩種方法皆為靜態系統的管制方式,忽略了變數之間的關連性和變數本身的序列關係,因此對於以傳統MPCA或MPLS做監控時,將有誤判發生。
本論文將從靜態批次統計程序控制方法,發展出動態批次統計程序控制方法,我們考慮利用傳統的多變數動態管制方法DPCA,將變數考慮為時間延遲項的線性組合,並且加上批次的概念,希望能夠將系統的批次、動態特性均加入考慮。除考慮以量測變數為主的批次監控外,利用PLS的概念,將產品品質的特性也納入考量,發展出一套真正合乎系統特性且又靈敏的監控技術。最後,再將此種方法推展至即時線上偵測,希望可以在系統異常發生之時甚至快要發生之時,就可以有效的發現系統異常,進而從事製程修正或停機的動作,以避免製程原料浪費或是更大危險的產生。這研究在過去文獻中從未被討論。最後利用模擬例子和工廠數據(特用化學品)做本理論的驗證。


MSPC methods in batch monitoring research only utilize MPCA and MPLS that are proposed by Nomikos and MacGregor (1995). They extended the original PCA and PLS concept to the finite in the batch process. MPCA is a modified version of the PCA technique for handling the experimental study by taking the batch data set with three-way arrays. The three-dimensional array (variables x time x batches) batch experimental data can be unfolded into two-dimensional matrix. Like PCA, MPCA can be employed to compress noisy and correlated measurements into a simpler and smaller informative subspace for measurement data sets that contain significant redundancies. MPLS, an extended version of MPCA, utilizes the data not only on the measured process variable trajectories but also on the final quality measurement at the end of each batch. It can improve the detectability of the process deterioration and remove the measurements that are irrelevant to the quality of the product. However, both methods are often applied on the statistic system data even though batch process information is dynamic. They will fail in detecting the occurrence of small disturbance because the variable will be not only interdependent between the different variables but also serially dependent within each variable series.
The methodology for the proposed on-line batch monitoring is derived from the traditional PCA with the lagged variable data and batch-to-batch variation. Like the counterpart of PCA, the concept of PLS will be applied when the process quality variables are available. Finally, the proposed method is extended into on-line predictive monitoring. The purpose of the method is not only to detect the process fault as early as possible but also to reduce the false alarm at the minimum. This is particularly important in the most industrial processes. This research direction has never been studied in other literature. Extensive testing on the simulation problems and the experimental plant data (from specialty chemicals) will be used here to show the effectiveness of the proposed method.


摘要……………………………………………………………………I
Abstract ………………………………………………………………II
目錄……………………………………………………………………IV
圖目錄…………………………………………………………………VI
第一章:前言………………………………………………………1
1-1文獻回顧…………………………………………………1
1-2研究動機…………………………………………………3
第二章:PCA為基礎之批次即時偵測…………………………… 4
2-1PCA……………………………………………………… 4
2-2MPCA管制……………………………………………… 7
2-2-1共變異係數矩陣討論……………………………………8
2-2-2MPCA即時偵測………………………………………… 9
2-2-3MPCA即時偵測的管制界線…………………………… 12
2-3BDPCA…………………………………………………… 14
2-3-1動態延遲項討論…………………………………………17
2-3-2共變異係數矩陣討論……………………………………19
2-3-3BDPCA即時偵測 ……………………………………… 21
2-3-4BDPCA即時偵測流程 ………………………………… 23
2-4討論………………………………………………………24
第三章:PLS為基礎之批次即時偵測…………………………… 32
3-1PLS ………………………………………………………32
3-2MPLS …………………………………………………… 36
3-2-1MPLS即時偵測 ………………………………………… 38
3-2-2MPLS即時偵測的管制界線 …………………………… 40
3-3BDPLS…………………………………………………… 41
3-3-1BDPLS即時偵測………………………………………… 46
3-3-2BDPLS即時偵測流程…………………………………… 47
3-4討論………………………………………………………49
第四章:結論………………………………………………………58
符號說明 ……………………………………………………………68
參考文獻 ……………………………………………………………70


1.Aparist F., “Hotelling’s T2 Control Chart with Adaptive Sample Sizes, ” Int. J. Prod. Res., 34, 10 (1996).2.Chen J., Liao C-M, Lin F. R. J., & Lu M-J, “Principal Component Analysis Based Control Charts with Memory Effect for Process Monitoring, ” Ind. Eng. Chem. Res., 40 (2001).3.Crowder S. V. and Hamilton M., “An EWMA for Monitoring a Process Standard Deviation, ” Journal of Quality Technology, 24 (1992).4.Fuchs C. and Benjamini Y., “Multivariate Profile Charts for Statistical Process Control, ” Technometrics., 36, 2 (1994).5.Gan F. F., “Joint Monitoring of Process Mean And Variance, ” Nonlinear Analysis, Theory, methods & Application., 30, 7 (1997).6.Jackson J. E. and Mudholkar G. S. Technometrics, 21 (1979)7.Hotelling, H. “Analysis of a Complex of Statistical Variables into Principal Components, ” J. Educ. Psychol., vol. 24, pp. 498-520, 1933.8.Kourti T.;Nomikos P. and MacGregor J. F., “Analysis, Monitoring and Fault Diagnosis of Batch Processes Using Multiblock and Multiway PLS,” J. Proc. Cont., 5. 4. (1995).9.Lowry C. A. and Montgomery D. C., “A Review of Multivariate Control Charts” IIE Transaction., 27 (1995).10.Mason R. L.;Tracy N. and Young J. C., “A Practical Approach for Interpreting Multivariate T2 Control Chart Signals, ” Journal of Quality Technology, 29, 4 (1997).11.Neogi D. and Schlas C. E., “Multivariate Statistical Analysis of an Emulsion Batch Process,” Ind. Eng. Chem. Res., 37 (1998).12.Nomikos P. and MacGregor J. F., “Monitoring Batch Processes Using Multiway Principal Component Analysis,” AIChE Journal, 40, 8 (1994).13.Nomikos P. and MacGregor J. F., “Multivariate SPC Charts for Monitoring Batch Processes,” Technometrics, 37, 1 (1995a).14.Nomikos P. and MacGregor J. F., “Multi-way Partial Least Squares in Monitoring Batch Pprocesses,” Chemometrics and Intelligent Laboratory Systems, 30 (1995b).15.Pearson, K. “On Lines and Planes of Closest Fit to System of Points in Space, ” Phil. Mag., vol. Ser. B, no. 2, pp. 559-572, 1901.16.Runger G. C. and Prabhu S. “A Markov Chain Model for the Multivariate exponentially Weighted Moving Averages Control Chart, ” Journal of the American Statistical Association, 91, 436 (1996).17.Shewhart, W. A., “Economic Control of Manufactured Product,” Van Nostrand Princeton, NJ (1931).18.Sullivan J. H. and Woodall W. H., “A Comparison of Multivariate Control Charts for Individual Observations, ” Journal of Quality Technology, 28, 4 (1996).19.Tracy N. D.;Young, J. C. and Mason, R. L., “Multivariate Control Charts for Individual Observations” Journal of Quality Technology, 24, 2 (1992).20.Wachs, A. and Lewin, D. R., “Improved PCA Methods for Process Disturbance and Failure Identification, ” AIChE Journal, 48, 8 (1999).21.Wenfu Ku;Robert H. Storer and Christos Georgakis, “Disturbance Detection and Isolation by Dynamic Principal Component Analysis, ” Chemometrics and Intelligent Laboratory Systems, 30 (1995).22.Wold S., “Exponentially Weighted Moving Principal Components Analysis Projections to Latent Structures, ” Chemometrics and Intelligent Laboratory Systems, 23 (1994). 23.Wold S.;Kettaneh N.;Friden H. and Holmberg A., “Modelling and Diagnostics of Batch Processes and Analogous Knetic Experiments,” Chemometrics and Intelligent Laboratory Systems, 44 (1998).

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