Compensator based improved model predictive control for Aero-engine

Abstract Model predictive control (MPC) can be applied to aero-engine for optimal command tracking and constraint handling. However, the performance of MPC is highly dependent on the accuracy of the predictive model. Therefore, a new structure MPC with an additional modeling error compensation loop is proposed, which can substantially reduce the dependence of MPC on the accuracy of the predictive model, and enhance the performance and real-time property of the system simultaneously. The new MPC replaces the traditional feedback correction loop with a compensation control loop to realize high-performance tracking over the large envelope with only one predictive model. The compensation loop controller employs an augmented discrete linear quadratic method for robust tracking capability. Simulations on the turboprop engine show that, compared with the standard MPC, the new MPC shows greater robustness over the envelope with smaller settling time, overshoot, and fluctuation of power turbine speed.