Research Summary for
- Grid 4 results of the RAE 2822 airfoil case with
Mach=0.30, Re=6.2x106 , and alpha=0.0
deg. have been obtained.
- In last week’s presentation, the discrepancy between
the CL values and the Cp distributions originated from the use of an extra
cell face outside airfoil surface in the integration of the forces. The CL
and CD values are corrected and given in Table 1.
- The grid convergence is better for the subsonic case (Figure 2) than the one observed for the transonic case.
Figure 1. Pressure distributions obtained with the Sp-Al model and the Min-Mod limiter for the RAE 2822 airfoil case with Mach=0.30, Re=6.2 x 10 6, and alpha=0.0 deg.
Table 1. CL and CD
values obtained for two RAE 2822 cases (Sp-Al model and Min-Mod limiter).
|
Case |
Grid level |
CL |
CD (drag
counts) |
|
|
Mach =0.3, Alpha=0.0 deg, Re=6.2x106 |
1 |
0.15940 |
191 |
|
|
2 |
0.19694 |
104 |
||
|
3 |
0.20546 |
85 |
||
|
4 |
0.20550 |
83 |
||
|
Mach =0.75, Alpha=3.19 deg, Re=6.2x106 |
1 |
0.68992 |
353 |
|
|
2 |
0.75094 |
298 |
||
|
3 |
0.77889 |
295 |
||
|
4 |
0.79341 |
302 |
||
Figure 2. CL vs. h (measure of grid spacing) obtained
with the Sp-Al model and the Min-Mod limiter for the RAE 2822 cases with Mach=0.30,
Re=6.2x10 6, alpha=0.0 deg, and Mach=0.75, Re=6.2x106,
alpha=3.19 deg.
- A subsonic inviscid
case (Mach=0.3) was run with and without using the Min-Mod limiter at grid
levels 1 and 2.
- Program versions (4.0.1
and 4.1.0) were also compared. For the inviscid runs, there was no
difference between the results of different versions.
- The cases run with and
without the limiter gave different CL values (Figure 3).
- L2 norm residual
convergence of the case without the limiter seems to be better compared to
the convergence of the case run with the limiter (Figure 4).
Figure 3. CL vs. cycle number obtained with and without
the limiters for the inviscid case (Mach=0.3 and alpha=0.0 deg) at grid levels
1 and 2.
Figure 4. L2 norm residual history obtained with and without the limiters for the inviscid case (Mach=0.3 and alpha=0.0 deg) at grid levels 1 and 2.
·
The effect of the program version is observed in viscous runs when the
different compressibility corrections are used as part of the turbulence
models.
Figure 5. CL vs. cycle number obtained with and without
different compressibility corrections used in different versions of GASP at
grid level 2.
·
The CL oscillations
observed in the transonic case with the k-w turbulence model can be due to the
CFL number (its magnitude and type: based on free-stream or local quantities)
and/or the coupling or uncoupling of turbulence equations in the
time-integration process. To be sure, I need to run a few more cases.
