Surrogate model based system identification and control of gas turbine engines

Ekinci, Sinan
The focus of this thesis is on small low-cost turbojet engines where the fuel flow is provided by a gear-type fuel pump rather than a more traditional fuel metering unit. The incorporation of such type of fuel flow actuation devices introduces additional nonlinearities into the modeling and control of the small turbojet engines. As an alternative solution, this thesis presents and applies online/offline parameter estimation and adaptive control methods to such sub-class turbojet engines. For the simulation and analysis, a high-fidelity turbojet engine model including the fuel system is developed. On the other hand, a nonlinear fast engine model is proposed to capture the shaft dynamics of a small turbojet engine equipped with a gear-type fuel pump. The parameters of the fast engine model are estimated using the regression analysis. The same identification procedure is also applied to real engine test data to verify the proposed approach. Next, an online parameter estimation method using the same fast engine model structure is applied to the aero-thermal turbojet engine model. Then, the outputs of the offline and online parameter estimation approaches are compared with each other. Later, the online parameter estimation algorithm is converted to an adaptive controller known as indirect-model reference adaptive control. Last, several control and failure scenarios are applied to the aero-thermal engine model to show its main advantages over the traditional control approaches and also to reveal its further application areas.
Citation Formats
S. Ekinci, “Surrogate model based system identification and control of gas turbine engines,” Ph.D. - Doctoral Program, Middle East Technical University, 2019.