Robust Control of a Gas Turbine with Wiener Model Uncertainty

The idea of robust control is employed to utilize a simpler Weiner model of a gas turbine instead of more complicated thermo dynamical models. Two sources of uncertainties are considered: (1) uncertainties due to nonlinear look-up tables in the Weiner model, which are obtained from experiments on the real engine, and (2) un-modeled dynamics which cannot be captured by the simple structure of the Wiener model. For simplicity, instead of the real engine experiments, a thermo dynamical non-linear model of the engine is considered. The frequency domain behavior of the Wiener model is compared with the nonlinear model and the observed deviations are considered as unknown uncertainties. Two robust controllers, using the well-known D-K iteration methods, are designed for the Wiener model under the presence of such uncertainties, which is then applied on the nonlinear model. Finally, for further investigations and introduction of more uncertainties, a real-time engine simulator is developed, which serves as a more uncertain system due to computational and communication time delays. Study of various time responses resulted by applying a set of step inputs with different magnitudes show that by design of a robust controller for simple Weiner's model of gas turbine, a better performance and robustness will be achieved in comparison with PID control methods. Besides, modeling and controller design have very simple structure in this method in contrast with complicated non-linear methods.