使用模糊控制理論發展多進多出批次控制器之研究

Development of Run-to-Run Controller for the Multiple-Input Multiple-Output System Using Fuzzy Control Theories

龔彬(Pin Kung)

指導教授 ： 范書愷

元智大學/工程學院/工業工程與管理學系/碩士(2004年)

摘要

批次控制在工業界實行多年，一直受到廣泛的使用，藉由整合統計製程管制及工程製程管制，使製程的變異能夠有效地減少。del Castillo (2002) 利用多變量指數加權移動平均控制器隨時對製程干擾的不良影響進行補償性的調整。
模糊控制器不同於傳統的多變量指數加權移動平均控制器，它是利用模糊集合理論的概念，建構歸屬函數，設計模糊邏輯控制器，達到優於傳統多變量指數加權移動平均控制器的控制效能。吾人以Min-max-gravity方法為基礎，以及利用Gupta, Kiszka and Trojan (1986)的多變量多進多出模糊系統，設計出模糊控制器。討論在多變量白噪音及IMA(1,1)的製程干擾模型影響下，透過這個模糊控制器來替代單一多變量指數加權移動平均控制器及雙重多變量指數加權移動平均控制器做回饋控制，其模擬結果發現相較於單一多變量指數加權移動平均控制器及雙重多變量指數加權移動平均控制器，模糊控制器能夠獲得較小且穩定的MSE統計量。

關鍵字

批次控制 ； 統計製程管制 ； 工程製程管制 ； 多變量指數加權移動平均控制器 ； 歸屬函數 ； 模糊控制器 ； 多變量多進多出模糊系統 ； 單一多變量指數加權移動平均控制器 ； 雙重多變量指數加權移動平均控制器

並列摘要

Run-to-Run control has been extensively applied in semiconductor industry. It has been recognized that these two techniques, statistical process control and engineering process control, can be integrated in prefect harmony to produce more effective tools for process variation reduction. del Castillo (2002) used multivariate EWMA controller to adjust and maintain the dynamic process output on target.
Multivariate fuzzy control is inherently different from conventional multivariate EWMA control. It is designed by fuzzy logic and set concept. As the proposed fuzzy control can structure an appropriate membership function, thus it can be shown more effective than conventional multivariate EWMA control. In this study, fuzzy logic is used to design the multivariate fuzzy controller for MIMO Run-to-Run applications based on min-max-gravity method under Gupta, Kiszka and Trojan (1986)’s MIMO fuzzy system. Under multivariate white noise and IMA (1,1) disturbances, in some specific circumstances the multivariate fuzzy controller can generate stabler control outputs with smaller MSEs than single-MEWMA and double-MEWMA controller.

並列關鍵字

Run-to-Run control ； statistical process control ； engineering process control ； multivariate EWMA controller ； membership function ； multivariate fuzzy control ； MIMO fuzzy system ； single-MEWMA controller ； double-MEWMA controller

參考文獻

Yeh, Jinn-Yi, “ Object-Oriented Design and Implementation of an Optimizing Run-to-Run Quality Controller for Semiconductor Manufacturing Process,’’ The University of Texas at Arlington, 1997.

Baaklini, N. and E. H. Mamdani, “Prescriptive methods for deriving control policy in a fuzzy-logic controller,” Electron. Lett., 11, 1975, 625-626.

Box, George E. P. and Alberto Luceño, “Discrete Proportional-Integral Control with Constrained Adjustment ,” The Statistician, 44, 4, 1995, 479-495.

Box, George and Tim Kramer, “Statistical Process Monitoring and Feedback Adjustment-A Discussion,” American Statistical Association and the American Society for Quality Technometrics, 34, 3, 1992, 251-285.

Capilla, Carmen, Alberto Ferrer and Rafael Romero, “Integration of Statistcal and Engineering Process Control in a Continuous Polymerization Process,” American Statistical Association and the American Society for Quality Technometrics, 41, 1, 1999, 14-28.