Approximate model predictive control for gas turbine engines

Abstract

A novel model predictive control strategy using instantaneous linearization of nonlinear models incorporating the generalized predictive control (GPC) called approximate model predictive control (AMPC) is used to control a shaft speed of a gas turbine engine. This method gives advantages over the nonlinear model predictive control (NMPC), which is computationally demanding and has local minimums. The performance of the model based control schemes is dependent on the accuracy of the process model, so firstly the paper examines the estimation of global nonlinear gas turbine models using NARMAX and neural network representations. The performance of the proposed methods is examined using a range of small and large random step tests. The results illustrate the improvements in control performance that can be achieved to that of gain-scheduling PID controllers.