- 學位論文
以模糊數學建模技術為基礎之智慧型熱軋鋼輥隙控制
Intelligent Gauge Control for Hot Rolling Mill with Fuzzy Modeling Technigue
若您是本文的作者，可授權文章由華藝線上圖書館中協助推廣。
摘要
鋼板熱軋的生產過程中，精軋區中的軋機常常因為更換不同軋延尺寸或不同材質的鋼板，導致預測輥隙設定值和軋延力值與實際軋延間隙及軋延力間的誤差過大，而增加軋壞的發生率。尤其以鋼帶頭端(head end)輥隙開度設定最難以預測與控制。本文針對鋼帶頭端輥隙開度的最佳化設定，提出以模糊理論為基礎，依據鋼帶頭端的軋延力預測誤差值，設計一智慧型控制器動態設定輥隙開度，使鋼帶能夠順利地穿帶，並且有效地提升鋼帶頭端厚度的精確度。 整個過程是屬於離線模擬，包含初始值設定模型、軋延數學模型和模糊控制器三部份。對於實際的軋延現象，以非線性關係式來建構軋延數學模型。至於模糊控制器方面，運用常態分布的方法對熱軋資料作處理並分類，建立各站的歸屬函數，根據模擬所得的頭端軋延力誤差，來判斷下站鋼帶頭端的厚度誤差，進而適當地調整輥隙，以縮小板厚誤差。最後根據中鋼公司提供的資料，評估整個動態輥隙設定系統的效能。模擬結果顯示，超過60%的資料，修正與改善其鋼帶頭端誤差，並且縮小約18%的板厚誤差。
並列摘要
In a hot rolling mill process, variations in strip thickness or material types could result in significant errors of setup values for finishing mill machines and increase the possibility of strip cobbling. It is difficult to predict the optimal setup values for all the rollers'' gap in finishing mills. This thesis focuses on the optimal roll gauge setup for the strip’s head end and then designs an intelligent controller to carry out the dynamic roll gauge setup system based on fuzzy control theory. According to the deviation between the predicted rolling forces and the actual rolling forces, in order to improve the accuracy of head end strip thickness and reduce the production cost. The total process belongs to off-line simulation, including an initial value setup model, a rolling mathematical model and a fuzzy controller. For the actual rolling phenomenon, we use the nonlinear equations to construct the rolling mathematical model. About the fuzzy controller, we adopt the possibility distribution method to deal with the rolling data and build the rolling stand’s membership function. According to the deviation between the predicted rolling forces and the simulated rolling forces, to estimate the strip head end thickness error of the next stand and adjust the roll gauge value accordingly to reduce the thickness error. Finally, we adopted on-line data provided by China Steel Corporation, to test the feasibility of the dynamic roll gauge setup system. From the simulation results, more than sixty percents of the strip head end thickness have better accuracy and their thickness errors reduce up to eighteen percents.
參考文獻
[1]Oda, T., et al., “Adaptive Technology for Thickness Control of Finisher Set-up on Hot Strip Mill”, ISIJ International, Vol.35, No.1, pp.42-49, 1995.
[2]Stephens, R. I. and Randall, A., “On-line adaptive control in the hot rolling of steel”, IEE Proc.-Control Theory Appl., Vol.144, No.1, pp.15-24, 1997.
[3]Takagi, T., and Sugeno, M., “Fuzzy Identification of Systems and Its Applications to Modeling and Control”, IEEE Trans. Syst. Man. Cybern., Vol. 15, No.1, pp.116-132, 1985.
[8]Yamada, F., et al., “Hot Strip Mill Mathematical Models and Set-Up Calculation”, IEEE Transaction on industry applications, Vol. 27, No.1, Jan/Feb 1991.
[9]Zarate, L. E. and Helman, H., “Determination of the thickness control parameters of the rolling process through the sensitivity method, using neural networks”, IEEE Intelligent Processing and Manufacturing of Materials, Vol. 1, pp.537-542, 1999.