Comparative study of numerical approaches to adaptive gas turbine cycle analysis

Significant increase in task complexity for modern gas-turbine propulsion systems drives the need for future advanced cycles' development. Further performance improvement can be achieved by increasing the number of engine controls. However, there is a lack of cycle analysis tools, suitable for the increased complexity of such engines. Towards bridging this gap, this work focuses on the computation time optimization of various mathematical approaches that could be implemented in future cycle-solving algorithms. At first, engine model is described as a set of engine variables and error functions, and is solved as an optimization problem. Then, the framework is updated to use advanced root-finding paradigms. Starting with Newton-Raphson, the model is improved by applying Broyden's and Miller's schemes and implementing solution existence validation. Finally, algorithms are compared in representative condition using increasingly complex turbojet and adaptive cycle turbofan configurations. As evaluation cases become more time consuming, associated time benefits also improve.