Nonlinear model predictive control of a laboratory gas turbine installation

The feasibility of model predictive control (MPC) applied to a laboratory gas turbine installation is investigated. MPC explicitly incorporates (input and output) constraints in its optimizations, which explains the choice for this computationally demanding control strategy. Strong nonlinearities, displayed by the gas turbine installation, cannot always be handled adequately by standard linear MPC. Therefore, we resort to nonlinear methods, based on successive linearization and nonlinear prediction as well as the combination of these. We implement these methods, using a nonlinear model of the installation, and compare them to linear MPC. It is shown that controller performance can be improved, without increasing controller execution-time excessively.