Please use this identifier to cite or link to this item: `http://hdl.handle.net/11455/2682`
 標題: 雙軸渦輪扇引擎之多變數控制器的設計和調整Design and Schedule of Multivariable Controllers for A Twin-Spool Turbofan Engine 作者: 何錫村He, Xi-Cun 關鍵字: Turbofan Engine;渦輪扇引擎;LQG/LTR;GA;ANFIS;LQG/LTR;遺傳演算法;適應性網路模糊推論 出版社: 機械工程學系 摘要: 本文主要是以JT3D型雙軸渦輪扇引擎為對象，以LQG/LTR法對一設計操作點設計其強健控制器，再將LQG/LTR控制器結合增益調節矩陣來改善其它操作點的性能表現。文中首先介紹JT3D引擎各主要構件特性及非線性數學模型，接著描述一般渦輪噴射引擎之控制需求，為了達到零穩態誤差，在線性模式內加入積分器成成擴增系統，再設計一LQG/LTR控制器以符合控制設計的考量。我們根據LQG/LTR法設計出的控制器具有不錯的命令執行和抵抗干擾的能力，並且對不同操作點的模式化誤差有不錯的強健性。接著設定增益調節矩陣，定義一性能指標將不同操作點和設計點間的系統迴路轉移函數矩陣最大奇異值的差異最小化，並利用遺傳演算法搜尋此性能指標以求得各增益調節點的增益矩陣。最後，利用適應性模糊網路推論完成全域增益矩陣的增益調節，並以電腦數值模擬來驗證其控制效果。In this paper, a gain scheduling technique is proposed with the LQG/LTR method to design a robust nonlinear controller for JT3D turbofan engine. A mathematical model of JT3D turbofan engine with the characteristics of its major components is first derived. A linearized model at a nominal operation point is then obtained to design a LQG/LTR controller such that desired performances of command following and disturbance rejection can be achieved. A scheduling matrix is further adopted and connected to the LQG/LTR controller to form a robust controller such that the desired performances can be maintained at other specified operation points. Gains of the scheduling matrix at these specified operation points are obtained by use of a genetic algorithm through minimizing a normalized loop transfer difference maximum singular value. To extend the robust controller to the whole operation range of the JT3D turbofan engine, the scheduling matrix is implemented as an adaptive-network-based fuzzy interference system (ANFIS). The ANFIS and the LQG/LTR controller form a nonlinear controller for the JT3D turbofan engine to achieve the desire performances at the operation points. Robust performances of the nonlinear controller at the operation range of the JT3D turbofan engine are validated through the computer simulations. URI: http://hdl.handle.net/11455/2682 Appears in Collections: 機械工程學系所

TAIR Related Article