DLR Contribution to the First High Lift Prediction Workshop

Kurzfassung

DLR’s contribution to the first AIAA High Lift Prediction Workshop (HiLiftPW-1) covers computations of all three scheduled test cases for the NASA trapezoidal wing in high lift configuration. The DLR finite volume code TAU has been employed as the flow solver. In a standard set-up the one-equation turbulence model of Spalart and Allmaras in the original formulation is used to model effects of turbulence. For selected grids and flow conditions, the k-ω SST model of Menter and a differential Reynolds stress model (SSG/LLR-ω ) developed by DLR have been considered. DLR contributed with two hybrid unstructured grid families to the workshop. The grids have been generated with the grid generation packages Centaur and Solar. A grid family with three Solar grids has been generated and provided to the workshop featuring grids of 12·10^6 , 37·10^6 , and 111·10^6 points for test case 1. In addition, a Solar grid of 37·10^6 points has been provided for test case 2, and a grid of 40·10^6 for the configuration including the slat and flap brackets (test case 3). DLR didn’t succeed in generating a fine-grid with the Centaur package. In order to complete a Centaur grid family with three grid levels an extra-coarse grid has been provided. Thus, the three levels of the Centaur grid family are realized by grids of 13·10^6 , 16·10^6 , and 32·10^6 points. In general a go o d agreement between the experimental evidence and the polar computations on the Solar and Centaur grids is found in terms of forces, moments and wing pressure distributions. The wing tip area with the rearward part of the main wing and the flap represents the most challenging part of the configuration, especially at angles of attack around maximum lift. The deviations between the TAU solutions and the experimental data in this area are only weakly influenced by the different grid topologies or turbulence models used. The influence of the grid resolution of both grid families is comparable, taking into account the different absolute resolution levels of both grid families. Including the slat and flap brackets leads to the expected lift decrease. Concerning the convergence properties, a strong dependence on the numerical start-up procedure has been detected in many of the computations at higher angles of attack.