| Summary: | 碩士 === 國立中興大學 === 機械工程研究所 === 82 ===
In this paper, a method is proposed to design a robust controller for F-100 turbofan engine by combining the features of LQG/LTR and LQG/RPS approaches. In the design process of the proposed method, the MIMO residue method for model reduction is adopted to reduce both the complexities in control system structure and controller dynamic order.
First, the model reduction technique is used for the F-100 engine to simflify a linear model of order 7. Since a zero-state error for step commands is desired, an augmented system is established by consisting of the reduced-order model and the integrators. Based on the augmented system, an LQG/LTR controller is then designed to meet the design considerations. The LQG/LTR method results controllers that have "good" command following and "good" disturbance rejection and are robust. with respect to modeling errors. Stability robustness of the control system with respect to some uncertain parameters is also anslyzed. Furthermore, the stability robustness of the designed LQG/LTR controller with respect to uncertain parameters is enhanced by using the LQG/RPS method developed in reference [9].
Finally, the orders of these two controllers are reduced for imple-mentation. The performances of these two controllers are validated and compared through the computer simulations. We also analyze stability region and discuss the result of performance simulation.
|
|---|
