A320 DDES on 2048 cores > Silvia Reuß, Dieter Schamborn > 30.04.2010
Transition locations on the LEISA high lift airfoilS.Reuß
A320 DDES on 2048 cores > Silvia Reuß, Dieter Schamborn > 30.04.2010
Slide 2
Available experimental data
Two different settings were measured:
3eOptV1 which was measured in the slotted test section of the low speed wind tunnel NWB. From this measurement infra red pictures as well as pressure distributions are available.
3eOptV2 which was measured in the closed test section of the NWB. From this measurement pressure distributions as well as accoustical measurements are available.
All grids are for the OptV2 geometry, but the difference between those two is minimal:
A320 DDES on 2048 cores > Silvia Reuß, Dieter Schamborn > 30.04.2010
Slide 3
Available experimental data
First we got the infrared measurement for the 3eOptV1 and the pressure distribution of the 3eOptV2 both for α=7° angle of attack. A comparison of the pressure distributions for the two different test sections (slotted/closed) revealed strong deviations (see next 3 slides)
After consultation with the experimentalists it was clear, that we needed a different angle of attack for comparison with the data from the slotted test section. The suggestion was to use the α=8° case. Also the suspicion arouse, that there might occur transition on the slat for this angle of attack.
Now that we received the infrared pictures for the other incidence angles, it is clear, that no transition should be found on the slat.
Since no evaluated data is available, we conclude by simple optical judgment, that the transition locations for the α= 7° and α=8° case do not change significantly
A320 DDES on 2048 cores > Silvia Reuß, Dieter Schamborn > 30.04.2010
Slide 4
Available experimental dataPressure distribution slat
Pressure distributions• OptV2 (closed section)• OptV1 (slotted section)
• Data is measured in three sections
A320 DDES on 2048 cores > Silvia Reuß, Dieter Schamborn > 30.04.2010
Slide 5
Available experimental dataPressure distribution wing
Pressure distributions• OptV2 (closed section)• OptV1 (slotted section)
• Data is measured in three sections
A320 DDES on 2048 cores > Silvia Reuß, Dieter Schamborn > 30.04.2010
Slide 6
Available experimental dataPressure distribution flap
Pressure distributions• OptV2 (closed section)• OptV1 (slotted section)
• Data is measured in three sections
Here the influence of the wind tunnel side walls can be clearly seen. The curves with the most points are the measurements at the mid section
A320 DDES on 2048 cores > Silvia Reuß, Dieter Schamborn > 30.04.2010
Slide 7
Flow
Available experimental dataIR slat and main wing upper side, α=7°
A320 DDES on 2048 cores > Silvia Reuß, Dieter Schamborn > 30.04.2010
Slide 8
Available experimental dataIR slat and main wing upper side, α=8° Flow
A320 DDES on 2048 cores > Silvia Reuß, Dieter Schamborn > 30.04.2010
Slide 9
Available experimental dataIR slat and main wing upper side, α=9° Flow
A320 DDES on 2048 cores > Silvia Reuß, Dieter Schamborn > 30.04.2010
Slide 10
Available experimental dataIR slat and main wing upper side, α=10° Flow
A320 DDES on 2048 cores > Silvia Reuß, Dieter Schamborn > 30.04.2010
Slide 11
Available experimental dataIR slat and main wing upper side, α=11° Flow
A320 DDES on 2048 cores > Silvia Reuß, Dieter Schamborn > 30.04.2010
Slide 12
Available experimental dataIR slat and main wing upper side, α=12° Flow
A320 DDES on 2048 cores > Silvia Reuß, Dieter Schamborn > 30.04.2010
Slide 13
Available experimental dataIR wing and flap upper side, α=7° Flow
A320 DDES on 2048 cores > Silvia Reuß, Dieter Schamborn > 30.04.2010
Slide 14
Available experimental dataIR wing and flap upper side, α=8° Flow
A320 DDES on 2048 cores > Silvia Reuß, Dieter Schamborn > 30.04.2010
Slide 15
Available experimental dataIR wing and flap upper side, α=9° Flow
A320 DDES on 2048 cores > Silvia Reuß, Dieter Schamborn > 30.04.2010
Slide 16
Available experimental dataIR wing and flap upper side, α=10° Flow
A320 DDES on 2048 cores > Silvia Reuß, Dieter Schamborn > 30.04.2010
Slide 17
Available experimental dataIR wing and flap upper side, α=11° Flow
A320 DDES on 2048 cores > Silvia Reuß, Dieter Schamborn > 30.04.2010
Slide 18
Available experimental dataIR wing and flap upper side, α=12° Flow
A320 DDES on 2048 cores > Silvia Reuß, Dieter Schamborn > 30.04.2010
Slide 19
New numerical resultsSpalart Allmaras Model A new grid was built with some modifications:
The farfield distance was increased to 100c
The resolution of the three element noses was reduced a bit
The resolution of the slat wake and above the flap was increased
The resulting grid has again about 200000 points per layer
Calculations with this new grid showed a clear difference compared to those on the old grid (2d/3d hybrid grid that can be found on the ATAAC site)
These differences are due to the small farfield distance! Calculations with farfield vortical correction show a clear trend towards the new results
New calculations use a critical N-factor of 7.18, where the theoretically expected value is in the range of 7.18 to 7.3. Originally this value should be calibrated using the experimentally given transition locations, but since all calculations showed earlier transition, this procedure was not successful.
A320 DDES on 2048 cores > Silvia Reuß, Dieter Schamborn > 30.04.2010
Slide 20
Pressure distribution comparison old/new gridSA model
A320 DDES on 2048 cores > Silvia Reuß, Dieter Schamborn > 30.04.2010
Slide 21
Pressure distribution new gridSA model
With the new grid a corrected angle of attack of α=5° is needed when the Spalar-Allmaras model is used
Even though the pressure distribution does not show the plateau on the flap the flow seperates
A320 DDES on 2048 cores > Silvia Reuß, Dieter Schamborn > 30.04.2010
Slide 22
Skin friction new gridSA model
With the new grid a corrected angle of attack of α=5° is needed when the Spalar-Allmaras model is used
Even though the pressure distribution does not show the plateau on the flap the flow seperates
A320 DDES on 2048 cores > Silvia Reuß, Dieter Schamborn > 30.04.2010
Slide 23
Convergence new gridSA model
The RANS calculations with the SA model converge, but slower as with the old grid, where about 40000 iterations were sufficient
A320 DDES on 2048 cores > Silvia Reuß, Dieter Schamborn > 30.04.2010
Slide 24
Pressure distribution new gridSST model
With the new grid a corrected angle of attack of α=6° is needed when the Menter-SST model is used
A320 DDES on 2048 cores > Silvia Reuß, Dieter Schamborn > 30.04.2010
Slide 25
Skin friction new gridSST model
With the new grid a corrected angle of attack of α=6° is needed when the Menter-SST model is used
A320 DDES on 2048 cores > Silvia Reuß, Dieter Schamborn > 30.04.2010
Slide 26
Convergence new gridSST model
The RANS calculations with the SST model do not converge in steady calculations
A320 DDES on 2048 cores > Silvia Reuß, Dieter Schamborn > 30.04.2010
Slide 27
Convergence new gridSST model
An unsteady restart from the steady solution yields a converged solution (time step is scaled for better presentability)
A320 DDES on 2048 cores > Silvia Reuß, Dieter Schamborn > 30.04.2010
Slide 28
Transition locations on new grid
The black lines indicate the old suggested transition locations. With the new grid and Ncrit=7.18 the SA model yields transition on the slat.
4° and 5° transition lines coincide
A320 DDES on 2048 cores > Silvia Reuß, Dieter Schamborn > 30.04.2010
Slide 29
New recommendations We recommend to use the new grid, to prevent wrong results because of the small
farfield distance (Can be found on the ATAAC site as hybrid_mandatory)
We recommend to use the SA model with a corrected angle of attack of α=5° and following transition locations:
We recommend to use the SST model with a corrected angle of attack of α=6° and following transition locations:
Since the SST model shows a much better agreement with the experimental data, DLR is considering to use SST based DES.
*) The transition location on the lower side of the wing did not converge completely, but is considered to have small influence. No experimental data is available for the lower side.
Slat Wing Flap
Upper side Laminar xtr=0.189 xtr=0.953
Lower side Laminar xtr=0.59-0.633* Laminar
Upper side Laminar xtr=0.1815 xtr=0.949
Lower side Laminar xtr=0.59-0.633* Laminar
A320 DDES on 2048 cores > Silvia Reuß, Dieter Schamborn > 30.04.2010
Slide 30
Wiggles in Pressure distribution
At several point some wiggles in the pressure distribution could be observed. A close look to the surfaces built by centaur reveals the reason: The surface normals at some cells deviate noticeably from the neighboring ones
I do not have an idea how to prevent centaur from producing such bad cells