No abstract

No abstract

No abstract

The flow at low Mach numbers and high angles of attack over axisymmetric configurations is not symmetric. The mechanism that triggers the asymmetry is a combination of a global (temporal) instability and a convective (spatial) instability. This latter instability is caused by roughness and other geometrical imperfections, which lead to roll angle dependent forces. The flow at these conditions has a complex vortex sheet structure, with two or three different flow regions. An accurate simulation by means of Computational Flow Dynamics (CFD) is thus very challenging, and many researchers have therefore employed Large Eddy Simulation (LES) codes. This study demonstrates that Unsteady Reynolds Averaged Navier-Stokes (URANS) methods are a suitable alternative, if Scale Adaptive Simulation (SAS) is used. This method is capable of capturing the main flow features, provided that fine meshes, which achieve geometrical similarity between the meshed geometry and the real object, and small-time steps are used. It is also demonstrated that, by using URANS methods in combination with SAS, strong differences in the global and local forces depending on the surface roughness of the model are obtained, a result which coincides with several wind tunnel tests.

Under a contract with AIRBUS MILITARY (AI-M), an exercise to analyze the potential of optimization techniques to improve the wing performances at cruise conditions has been carried out by using an in-house design code. The original wing was provided by AI-M and several constraints were posed for the redesign. To maximize the aerodynamic e©ciency at cruise, optimizations were performed using the design techniques developed internally at INTA under a research program (Programa de Termo §uidodin‚ amica). The code is a gradient-based optimization code, which uses classical ¦nite di¨erences approach for gradient computations. Several techniques for search direction computation are implemented for unconstrained and constrained problems. Techniques for geometry modi¦cations are based on di¨erent approaches which include perturbation functions for the thickness and/or mean line distributions and others by B‚ ezier curves ¦tting of certain degree. It is very important to a¨ord a real design which involves several constraints that reduce signi¦cantly the feasible design space. And the assessment of the code is needed in order to check the capabilities and the possible drawbacks. Lessons learnt will help in the development of future enhancements. In addition, the validation of the results was done using also the well-known T AU §ow solver and a far-¦eld drag method in order to determine accurately the improvement in terms of drag counts.