Study Of Predictive Control For A Nonlinear Aircraft System

Since introduced by Norbert Wiener (1894-1964), control theory has endured a long period development. In the end of 70's,coming up with the development of digital computer to minitype, high speed, huge capacity and low-cost, predictive control was born.Predictive control, which has difference at the initial point of theory from PID control theory, is not an outlet from any of other theories, but developed from industrial reality. PID control theory decides the controller generally based on past information and the set point of the output. However, predictive control, not only uses the past and the present information, but also uses the future information gained by predictive model, and then decides the best control strategy based on incursion of step. Predictive control is prior to PID control theory from the point of idealism. This article is mainly about the Generalized Predictive Control (GPC) theory, which brought by CLARK in 1987. As a kind of self-tuning control algorithm, GPC is used in discrete system. Recent years, GPC attracted a lot of attention, brought many reference and papers, and got lots of successful applications in control systems of some industrial fields such as Petroleum, Chemical, Power, Metallurgical, Mechanical, etc.With GPC theory, an aircraft dynamical model is studied in this thesis. Firstly, a dynamic model of an aircraft is developed. This model is a nonlinear state-space model with four input variables, four output variables and nine state variables. Secondly, from single input single output model to two inputs two outputs model, and then to some relatively complex linear model, the application of GPC theory is studied. Nonminimum-phase process is also studied in the field of linear and nonlinear system theory by model-algorithmic control. At last, a nonlinear aircraft model is studied with GPC controller. A stable control system is realized, and pleased simulation result is given in the end of this article.The major contribution of this thesis is the development and implementation of GPC control algorithms and realization of a GPC program which is able to regulate different linear multi-input multi-output processes. Nonminimum-phase system, linear system with white noise or with big time-delay, all can be applied to this GPC program. The most difficult problem solved is the access of stable control result, for the complexity of the nonlinear model. A motor model, a water tank model and a chemical reactor model are also studied, and pleased simulation results are accessed.