透過您的圖書館登入
IP:3.141.30.211
  • 學位論文

鋼板熱軋製程中溫度控制系統之圖控式模擬器及前饋控制律之設計

Developing Simulator for Hot Rolling during Finishing Mill Process and FeedForward Controller Law

指導教授 : 吳昌暉
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


本論文所提之鋼板製程中精軋區圖控式模擬系統,係以建立在精確的數值式精軋區溫度模型為基礎。此動態模擬系統,根據入口溫度、速度變化、及灑水量變化等線上資料,可以以離線的方式重現熱軋過程中的溫度變化情形,且亦可與數值式熱軋模型的預測結果吻合,使完軋溫度估測誤差之均方根值小於5C。 基於圖控式模擬器的開發成功,得以進一步研究前饋控制器於消弭鋼板入口溫度平均的低溫痕低溫痕;本文針對低溫痕的問題,提出兩種前饋控制器的設計概念,DM_FFC(Disturbance Measurement FeedForward Controller)及DT_FFC ( Dynamic Target_Temperature FeedForward Controller),分別以不同的方式來取得低溫痕的訊息,配合P Gain和反水濂函數來解決低溫痕的問題。兩個不同架構的前饋控制器均能達到95%消弭低溫痕的目的。 為了讓前饋增益值(FeedForward Gain)得以自我調適,以因應參數估測誤差的可能狀況,而導入相關係數的概念,並求得相關係數與增益值調適變化的關係式,使得本系統得以經過調適的過程得到一適當的增益值;由於本前饋控制器的發展是基於圖控式模擬器出發的,因此,為了解模擬器的參數偏差對於前饋控制器的影響,於文末亦做風險評估的測試,評估的結果為20%的參數偏差,前饋控制器仍可以維持消弭低溫痕達90%的水準。

並列摘要


This paper presents a study for developing a simulation system for hot rolling during finishing mill process based on graphic user interface. The simulation system is developed from a numerical exactly thermal model. According to the exit temperature , velocity data, and quantity of water spray, or water curtain, this dynamic simulation system can be applied to calculate the changing of temperature during the finishing milling process by off line mode; As the result agrees to the one calculated by numerical thermal model. That is the RMS error is lower than 5C. Resulting from the success of simulator development, this project can be moved forward to the development of feedforward controller for canceling the influence of skin mark. This study mentions two feedforward control schemes for skin mark problem. That is Disturbance Measurement FeedForward Controller(DM_FFC), and Dynamic Target_Temperature FeedForward Controller(DT_FFC). By using of the different processes of extracting skin mark information from temperature data, the skin mark can be completely canceled with suitable feedforward gain and the inverse of non-linear water efficiency function. Both of the above schemes for feedforward controller can be utilized to achieve the goal to cancel the skin mark by 95%. The correlation coefficient is introduced for feedward gain self-tuning. From the correlation coefficient and the variation of feedforward gain, the relationship function is obtained for feedforward gain tuning. Since those feedforward controller schemes are developed from the simulation system, based on the numerical thermal model, the influence of parameter deviations on feedforward controllers is regarding as a crucial information. At the end of this study, the evaluation for robust test is conducted. Resulting as, the feedforward controllers can still be working fine for canceling the skin mark by 90%, when there is +20% parameter deviations of thermal model.

參考文獻


[2] 王敏和,"熱軋廠精軋區溫控系統前饋與回饋控制器設計",私立元智工學院,碩士論文,民國86年。
[3] 武宛君,”鋼板製程中精軋區熱模型之建立”,私立元智大學,碩士論文,民國88年。
[5] F.Yamada,et al.,”Hot Strip Mill Mathematical Models and Set-up Calculation”,IEEE Transactions on Industry Applications,27,1,pp.131-139.January/February 1991.
[6] S.ZUNFT,”Temperature Control of A Distributed Collector Field”,Solar Energy, pp 321-325, No. 4 vol.55, 1995.
[8] Sadasiva, I.,et al. “A Graphical Based Automatic Real Time Code Generator for Power Electronic Control Applications”,Industrial Electronics, 1997.ISIE’97.,Proceedings of the IEEE International Symposium on, pp942-947 vol.3, 1997

延伸閱讀