Nonlinear model predictive control for thrust tracking of a gas turbine

In this paper, we investigate application of Nonlinear Model Predictive Control (NMPC) for controlling thrust of a single spool gas turbine engine. We showcase the development of a nonlinear black-box model based on Weiner modeling approach. The model consists of linear dynamic block parameterized by Laguerre, orthonormal basis functions, followed by a static nonlinear block captured in Artificial Neural Network (ANN). This model is used to predict gas turbine performance and a closed-loop controller is realized using NMPC formulation. Model validation in open-loop highlights potential of the identification methodology. Closed-loop simulation results showcase very satisfactory control performance for gas turbine engine thrust tracking control problem and bring forth the applicability of NARC formulation for highly nonlinear systems with fast dynamics.