Design of QFT/H∞ Composite Controller for Uncertain Systems

• Main Authors: Chung-KuangChen, 陳重光
• Other Authors: Cheng-Neng Huang
• Format: Others
• Language: en_US
• Published: 2015
• Online Access: http://ndltd.ncl.edu.tw/handle/79014552900618902676

碩士 === 國立成功大學 === 系統及船舶機電工程學系 === 103 === For the complicated systems, the plant parameter uncertainties and external disturbances will lead to unfavorable performance or system instability. For this reason, the Quantitative Feedback Theory (QFT) is explored in this research to offer optimal weighting functions to the standard H∞ control problem so that the system robustness and performance can be maintained. The lag-lead compensator and the servo-compensator will first be formulated into the system so that the desired specifications and perfect tracking can be achieved by adjusting the control parameters in the lag-lead control component. Furthermore, the weighting functions, which ensure the closed-loop stability under plant uncertainties, are then augmented to the system to form the standard H∞ control problem in order to get the H∞ optimal control gains. To select weighting functions, the sensitivity function S(s) and the complementary function T_1 (s) under plant parameter variation are analyzed in the frequency spectrum such that the satisfaction of the set of inequalities proposed in this research will guarantee the system robustness. Moreover, the optimal H∞ controller will eliminate the harmful influence of the exogenous inputs (i.e., disturbances and noises) on the controlled outputs (i.e., tracking errors and the control energy) so as to improve the robustness and tracking performance of the system. As for the uncancelled nonlinearities and uncertainties in the plant, the Popov stability criterion is then utilized to ensure the closed-loop stability. Finally, two examples with uncertainties are simulated to illustrate the achievement of tracking performance of the systems. The computer simulation results show that these systems can meet their desired specifications even though these systems encounter external disturbances and parameter uncertainties.