Optimisation du contrôle actif de la couche limité par aspiration sur un profil

The primary objective of this thesis is the development of an optimization code for the active control of the boundary layer by suction. Suction changes the structure of the boundary layer by reducing its thickness and increasing its stability. This has two essential effects on the flow: (1) delay of the laminar-turbulent transition; (2) modification of the intensity and position of the shock in transonic flows.

For low rates of suction of ten cm³/ s (for 1 m of scale), the impact is significant on the aerodynamic performances of a profile.

The large computing time for the prediction of the transition point has compelled us to carryout the calculations by using three coupled codes (a code for solving the Euler equations, a code for boundary layer correction and a code for the prediction of the transition). In order to use them within an algorithm of optimization, these three codes must be able to communicate automatically with each other.

The proposed method of transition prediction uses the stability theory. We used the SCOLIC code, developed at the Bombardier Chair, which solves the Orr-Sommerfeld equations.

In our concern to improve the quality of the results, we have also worked at the development of a code for the predication of the transition using the compressible Parabolized Stability Equations (PSE).

For the calculation of the flow we have modified an Euler code, developed at the Chair, in order to couple it to a boundary layer code (CLDF-ONERA). Finally we have developed an original method for the optimization of the suction, coupled with the results of the codes previously written. For that we have used Kriging and a learning algorithm.

The present thesis summarizes the theories and the algorithms set up for this optimization. Finally, we present the results of this optimization on a laminar profile.