Summary of Data from the Sixth AIAA CFD Drag Prediction Workshop: CRM Cases 2 to 5

Abstract

Results from the Sixth AIAA CFD Drag Prediction Workshop Cases 2 to 5 are presented. These cases focused on force/moment and pressure predictions for the NASA Common Research Model wing-body and wing-body-nacellepylon configurations. The Common Research Model geometry differed from previous workshops in that it was deformed to the appropriate static aeroelastic twist and deflection at each specified angle of attack. The grid refinement study and nacelle-pylon drag increment prediction (Case 2) used a common set of overset and unstructured grids, as well as user-created multiblock structured, unstructured, and Cartesian-based grids. Solutions were requested for both the wing-body and wing-body-nacelle-pylon at a fixed Mach number and lift coefficient. The wing-body static aeroelastic/buffet study (Case 3) specified an angle-of-attack sweep at finely spaced intervals through the zone where wing separation was expected to begin. The optional Case 4 requested grid adaption solutions of the wing-body at a specified flight condition. Optional Case 5 requested coupled aerostructural wing-body solutions. Results from this workshop highlight the progress made since the last workshop, and the continuing need for computational fluid dynamics (CFD) improvement, particularly for conditions with significant flow separation. These comparisons also suggest the need for improved experimental diagnostics to guide future CFD development.

Contributions to HiLiftPW-3 Using Structured, Overset Grid Methods

“…The effect of using Quadratic-Constitutive-Relation (QCR) [21] is well documented in various paper and over the last few DPW workshops [10][11][12]. Typically, the use of QCR results in improved results in corner flow separated regions.…”

“…The development of the HL-CRM was motivated by the success of the Common Research Model (CRM) [9]. The CRM has become a standard for both the experimental and CFD modeling community where it represents a public domain physically consistent geometry which has been used in numerous workshops (the DPW series [10][11][12], high-order workshop series and here the HiLiftPW3 [5]) and is currently a standard for a number of experiments across the international community. The HL-CRM was developed by Boeing (see, Ref.…”

Winglet Geometry Impact on DLR-F4 Aerodynamics and an Analysis of a Hyperbolic Winglet Concept

“…The impact of different cant angles for a classical Whitcomb winglet was identified in Mach number range of M = 0.6~0.87, with Reynolds' number being equal to As a reference area for calculating the aerodynamic coefficients of lift and drag, the wing plan area was taken including its ventral (non-wetted) part S = 0.148 m 2 . The dimensions of the simulated domain and the mesh resolution were chosen to match the materials presented at AIAA CFD Drag Predictions (DP) seminars, for instance [30,31]. The simulated domain dimensions were five body lengths along the symmetry axis of the model, 2.6 body lengths along the vertical axis and 2.5 body lengths along the z-axis.…”

“…An example of a typical grid used in stress calculations is shown in Figure 7. (Germany), ONERA (France), DRA (UK) and National Aerospace Laboratory of the Netherlands [30,31], allowed a verification of the robustness of our research technique, and validate the selected boundary conditions, mesh resolution, and turbulence models for the Mach and Reynold's numbers' ranges corresponding to typical climb and cruising flight mode conditions, a comparison of ANSYS Fluent computational results using different turbulence models with selected AGARD experimental results from [33,34] is given at Figure 6 below for the lift and drag coefficients at flight angles of attack.…”

Nonlinear corrector for Reynolds‐averaged Navier‐Stokes equations

“…Mesh convergence study is thus mandatory to have a reasonable confidence into the discrete outputs, but it may fail. For instance, the AIAA CFD prediction workshops [5][6][7][8] have pointed out the mesh dependency on the obtained results; in other words, two family of meshes lead to two different answers. In consequence, very strict meshing guidelines based on previous experiences * are provided to minimize this *It means that somehow we already know the solution!…”