Unsteady optimal flow control on aerodynamic applications
Final Report Abstract
During the landing approach of an aircraft, retractable flaps and slats are extended to increase the lift of the main wing. It allows the aircraft to slow down which is essential for e.g. a safe landing. But such high lift configurations are complex and heavy and lead to an overall less efficient aircraft. The size of e.g. a flap could be reduced if the unfavorable separation could be delayed to higher angles of attack. The present project was investigating the use of synthetic jets in the flap region to achieve such a delay. The main focus was how to find the right locations and the right parameters for such an active flow control method featuring a tremendous amount of parameters. The developed and investigated method is an adjoint flow solver that allows the efficient optimization of active flow control actuators on a high-lift configuration. In contrast to standard optimization procedures, the adjoint flow solver allows to calculate the sensitivities of an unlimited number of parameters at a fixed cost. It is therefore suitable for applications with a huge amount of design parameters. Two different methods to generate an adjoint flow solver were compared based on the same flow solver; the continuous adjoint and the discrete adjoint. For the initial main questions, satisfying answers could obtained: How does the continuous and the discrete approach based on the same flow solver compare to each other in the context of key features like accuracy, robustness and computational efficiency? - The developed discrete adjoint solver successfully calculated the sensitivities of all test cases with excellent precision. The accuracy was also maintained on coarse grids where the continuous solver started to degrade. Is it possible to improve the computational efficiency of the discrete solver to compete with the continuous? - Comparing the runtime of a full optimization cycle, the highly optimized discrete adjoint solver requires approximately twice the time of the continuous adjoint solver. This penalty is surprisingly small considering the advantages in terms of accuracy and robustness of the adjoint solver. Is it possible to obtain accurate sensitivities for an unsteady RANS simulation of a standard three element high-lift configuration? - Nearly perfect accuracy could be demonstrated. The full optimization of a high-lift configuration’s flap equipped with synthetic jets lead to a lift increase of 29%.
Publications
- Optimal control of unsteady flows using discrete adjoints. AIAA paper 2011-3720, 2011
Anil Nemili, Emre Özkaya, Nicolas R. Gauger, Angelo Carnarius and Frank Thiele
- A discrete adjoint approach for unsteady optimal flow control. Notes on Numerical Fluid Mechanics and Multidisciplinary Design, 121:447-455, 2013
Anil Nemili, Emre Özkaya, Nicolas R. Gauger, Angelo Carnarius and Frank Thiele
- Discrete adjoint based sensitivity analysis for optimal flow control of a 3D high-lift configuration. AIAA paper 2013-2585, 2013
Anil Nemili, Emre Özkaya, Nicolas R. Gauger, Felix Kramer, Angelo Carnarius and Frank Thiele
- Optimal control of unsteady flows using a discrete and a continuous adjoint approach. System Modeling and Optimization. IFIP Advances in Information and Communication Technology, 391:318-327, 2013
Angelo Carnarius, Frank Thiele, Emre Özkaya, Anil Nemili and Nicolas R. Gauger
- Optimal Design of active flow control for a complex high-lift configuration. AIAA paper 2014-2515, 2014.
Anil Nemili, Emre Özkaya, Nicolas R. Gauger, Felix Kramer, Tobias Höll and Frank Thiele
(See online at https://doi.org/10.2514/6.2014-2515) - Optimal separation control on the flap of a 2D high-lift configuration. Computational Methods in Applied Sciences, 36:411-425, 2015
Anil Nemili, Emre Özkaya, Nicolas R. Gauger, Felix Kramer and Frank Thiele
(See online at https://doi.org/10.1007/978-3-319-11541-2_27)
