- I have started running a sample case for the airfoil problem. This is the RAE 2822, Mach=0.725, Re_c=6.5 x 106 , alpha=2.92 deg. case taken from the AGARD report summarized last week. I got the C-grid (369x65x2) from the NPARC validation archive. The same case was run with the same grid in AeroSoft for GASPv4 validation. The purpose of these initial runs were to have some experience for the airfoil problems in GASP before starting an intensive study. Three grid levels were used in the runs: fine (369x65x2), medium (185x33x2), and coarse (93x17x2). For the fine grid, 304 points were located on the airfoil. The outer boundary of the mesh is 27 chord lengths away from the airfoil surface. Figure 1 and 2 show the fine grid used in the computations.
- Runs were performed with 3rd order spatially accurate upwind-biased Roe flux scheme, Sp-Al turbulence model, and the Min-Mod limiter. All the viscous terms were included in the computations. Whole flow field is assumed to be turbulent.
- Runs were performed for two angles of attack (AoA): 2.92 deg. used in the experiments (no correction) and 2.40 deg. used by Coakley (from Viscous Transonic Airfoil Workshop results). Different results are given in Figures 3 to 8.
- I had a meeting with Dr. Neel in AeroSoft. Some of the issues we discussed were:
- The case we'll choose for our airfoil problem
- Grid quality for airfoil problems
- Effect of using different boundary conditions for the far-field surfaces on the results.
- Effect of the distance from the airfoil surface to the outer boundary (GASP does nat have a circulation correction option for the outer boundary)
- Modeling the transition location in GASP
Figure 1. Fine gird (368 x 64 cells) used in the RAE 2822 airfoil computations.
Figure 2. Close up view of the fine gird (368 x 64 cells) used in the RAE 2822 airfoil computations.
Figure 3. Pressure distribution on the RAE 2822 airfoil obtained
with different grid levels, alpha=2.92 deg. (uncorrected).
Figure 4. Pressure distribution on the RAE 2822 airfoil obtained
with different grid levels, alpha=2.40 deg. (Coakley's correction).
Table 1. CL comparison with the experiment at different grid levels.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Table 2. CD (in drag counts) comparison with the experiment at different grid levels.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Figure 5. CL vs. cycle number for different grid levels, alpha=2.92
deg. (uncorrected).
Figure 6. CL vs. cycle number for different grid levels, alpha=2.40 deg. (Coakley's correction).
Figure 7. L2 norm residual history for different grid levels, alpha=2.92 deg. (uncorrected).
Figure 8. L2 norm residual history for different grid levels,
alpha=2.40 deg. (Coakley's correction).