Aero-engine direct thrust control based on nonlinear model predictive control with composite predictive model

Abstract

Abstract A novel NMPC (Nonlinear Model Predictive Control) based on composite predictive model is proposed and applied to direct engine thrust control. To improve the real-time of NMPC, an adaptive composite model based on SVM (State Variable Model), KF (Kalman Filter), and CLM (Component Level Model) is proposed as predictive model. The correction theory is adopted to establish a full envelope adaptive on-board predictive dynamic model and reduce the data storage of predictive model. At each sampling time, the CLM is calculated only once in the proposed NMPC, instead of many times in the popular NMPC based on EKF (extended Kalman filler). Therefore, the proposed NMPC has better real-time performance than the popular one. The simulations that consist of the proposed NMPC, the popular NMPC based on EKF, and the traditional controller PID are conducted. The simulations demonstrate that the proposed NMPC not only has greatly better real time performance than popular NMPC, but also has faster response speed than traditional controller PID.