See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/268558066

Flow Prediction around the SACCON Configuration Using CFD++

Conference Paper · June 2010

DOI: 10.2514/6.2010-4563

CITATIONS READS
3 380

3 authors, including:

U. Goldberg
Metacomp Technologies, Inc.
96 PUBLICATIONS 2,432 CITATIONS

SEE PROFILE

All content following this page was uploaded by U. Goldberg on 16 November 2016.

The user has requested enhancement of the downloaded file.

 Flow Prediction around the SACCON Configuration Using
CFD++
David Chi, Sukumar Chakravarthy and Uri Goldberg
Metacomp Technologies, Inc.
Agoura Hills, CA

Abstract
A three-dimensional numerical simulation of a vortex breakdown, using the UCAV half
wing configuration, has been performed. The vortex-dominated flow-field, where non-linear
effects have a significant impact over the wing, causes loss of lift. These nonlinear regimes
are the areas where linear stability and control methods normally fail. These tests, using
CFD++, were performed primarily at the low-speed side of the static stability and control
characteristics of this configuration. Two turbulence models, used for validation purposes,
were the SA (Spalart-Allmaras) and kω-SST(Shear-Stress Transport) closures to capture
these complex flow fields at angle-of-attack ranges of 15° to 22°. Detailed results compare the
numerical to experimental data in terms of pressure coefficient, vorticity components, lift,
drag and pitching moment coefficient (Cl, Cmy, Cd). The numerical results indicate that the
one-equation turbulence model (Spalart-Allmaras) yields better moment predictions than
those by the two-equation model (SST), but the latter predicts lift and drag better.

I. Nomenclature

Vµ = Free stream velocity [m/s]

n = Kinematic viscosity
F = Reference area
x, y, z = model coordinates in [mm]
cp = Pressure coefficient
s = Half span
MRP = Moment Reference Point
cr = Root chord length of the model
x/cr = Non-dimensional chord-wise cooridnate
cref = Reference length
cp = (p-p¥)/q¥ = Pressure coefficient
SA = Spalart-Allmaras turb. Model
SST = Menter Shear Stress Transport turb. Model
cMy = M/(q¥×F× cref )= Pitching moment coefficient
AoA, a = Angle-of- attack (nose up positive)
cL = L /(q¥××F) = Lift coefficient
cd= l /(q¥××F) = Drag coefficient
M = Mach number
Re = Vµ× cref /n = Reynolds number
q¥ = r¥ ×V¥2/2 = Dynamic pressure coefficient
y = Span-wise coordinate (m)
y/s = Non-dimensional span-wise coordinate

1
 II. Introduction

Understanding and predicting static stability of the Unmanned Combat Air Vehicle (UCAV) configuration at
various AoA are important factors in the vehicle design process. UCAV is designed using a delta wing planform
configuration with a 53° leading-edge sweep. One advantage of using a delta wing planform is that, as the angle- of-
attack increases, the leading edge of the wing generates a vortex which remains attached to the upper surface of the
wing, giving it a very high stall angle, thus maintaining lift at high AoAs. This vortex reduces upper surface pressure
by inducing high velocities on the upper surface.

The flow over delta wings is dominated by leading edge vortices and other vortical structures. UCAV relies on
lift generated by the vortex flow to enhance its stability and control. At a high angle-of-attack, strong vortices are
generated along the leading edge of the highly swept delta wing. These leading-edge vortices have a dominant effect
on the flow and aerodynamic loads. At such high angle-of-attack, a phenomenon known as “vortex breakdown”
occurs in the vortices above the wing. When breakdown occurs over a wing surface, it causes a sudden lift loss and a
severe non-linearity in the aerodynamic loads.

For many years, several researchers have investigated experimentally1,2,3 and numerically4,5,6,7 these vortex
phenomenon for developing the UCAV configuration. Design studies were also done on the effects of vortices7
around the wing. Ghaffai8 carried out experimental and numerical studies on the delta wing and investigated the
formation of leading edge vortices, which increase the maximum lift coefficient by adding vortex-induced lift.

One of the most important studies includes the effect of geometrical parameters such as sweep and leading edge
radius, as well as flow parameters like Mach number, Reynolds number and angle-of-attack. Hitzel9 tested a medium
sweep wing of 53° with two outer swing angles of 53° and 65° and three leading edge shapes (round, mid-round and
sharp). To validate the numerical simulations, several experimental tests were performed. An overview of static test
results is given by Loeser et al.10 and PIV data are presented by Gilliot11 and Konrath et al.12

The objective of this paper is to provide understanding of the flow physics as well as best practice methods for
computational simulation of vortex-dominated flow-fields where non-linear effects have a significant impact. Two
turbulence models, SA and SST, were used to study modeling effects on prediction of pressure distribution, lift, drag,
and pitching moment coefficients. Experimental data contain static measurement of integral forces and moments,
pressure distributions over the wing surface as well as field measurements using PIV – Particle Image Velocimetry.
A combination of the numerical results and experimental data leads to a better understanding of the complex flow
structure.

III. Methodology

A. Grid

For the current simulations, half the wing was meshed with a symmetry plane and no sting. In total, the flow
domain around half-span configuration contains approximately 12 million unstructured grid cells (Tetrahedral,
Triangular prisms, pyramids, and Hexahedral elements). Figures 1 and 2 show the partial computational grid around
the wing and in the boundary layer region of the delta wing.

2
 Figure 1. Grid around wing Figure 2. Grid in boundary layer region

B. Solver

CFD++15, a general Navier-Stokes flow solver, was used here. This code features a second order Total Variation
Diminishing (TVD) discretization based on a multi-dimensional interpolation framework. For the results presented
here, an HLLC (Harten, Lax, Van Leer, with Contact wave) Riemann solver was used to define the (slope-limited)
upwind fluxes. This Riemann solver is particularly suitable for high-speed flow applications since, unlike classical
linear solvers such as Roe’s scheme, it automatically enforces entropy and positively conditions16. Further details on
the numerical methodology in CFD++ can be found in Chakravathy et al17, and Peroominal el al18,19.

The preconditioned form of the compressible Navier-Stokes equations was used in the current effort. Turbulence
closure was achieved by use of either the SA or the SST turbulence model. The simulations were performed in
steady state mode. Stoke’s hypothesis for a Newtonian fluid and the Sutherland’s law for dynamic molecular
viscosity were assumed to apply throughout the flow-field.

C. Boundary and initial conditions

Pressure and temperature were set to 101325 Pa and 295 K, respectively. Walls were set to No-slip and adiabatic
boundaries. Since the outer boundary surface was placed relatively close to the wing, absorbing far-field layers were
applied to the cells adjacent to the outer boundary.

D.Turbulence models13,14

Two turbulence models, Spalart Allmaras and Menter SST, were used. Both models are based Boussinesq
hypothesis for turbulence production. The SA model is a turbulence closure comprising a transport equation for the
eddy viscosity. The SST k-ω turbulence model is a two-equation eddy-viscosity closure comprising transport
equations for the turbulence kinetic energy, k, and the turbulence inverse time scale, .

E. UCAV configuration

The simulated configuration comprises a delta wing planform with a 53° leading edge sweep and a twisted
lambda configuration. Figure 3 shows the planform and reference data for the SACCON configuration. The three
main sections of the model are the fuselage, wing panel section and wing tip. The outer wing section is twisted by 5°
around the leading edge to reduce aerodynamics load. The reference root cord is approximately 1m long from apex

3
and the wing span is 1.54m (0.77m for half-wing from the center). The reference length (cref =0.479 m) is used in Re
and moment calculations.

Figure 4 shows the position of the pressure taps on the SACCON model. Chord-wise positions of the
measurement planes are at 20%, 45% and 70%. The 20% location (non-dimensional chord wise) corresponds to
flow transition at the apex. The 45% location corresponds to flow transition between fuselade and inner wing. The
70% location is the vortex breakdown zone. Two additional span-wise positions, at y/s = 62% and 94%, contained
cross wing measurement data.

Figure 3. Planform and geometric parameters Figure 4. Position of pressure taps on the
of the UCAV SACCON configuration7 SACCON model

IV. Numerical Simulation

A 53° swept and twisted lambda wing is considered in this simulation. The simulations were done on a 12
million cell unstructured mesh with two different turbulence models, SA and SST models. Half wing configuration
used in the simulations is shown in Fig. 4. In this section, the results of the different computational methods will be
discussed and compared to experimental data. In all computations, fully turbulent flow conditions were assumed.
Free stream conditions of the wind tunnel are M = 0.146 and Re = 1.6×106. The Y+ level of the first grid layer of the
wing was controlled to be close to 1.

In Figures 5 to 7, numerical results using the CFD++ solver are presented for the lift, pitching moment, and drag
coefficients versus angle-of-attack in the range 15° to 22°, compared to the experimental data. In Figure 5, lift

4
coefficient predictions are below the experimental data till 19°. Figure 6 shows that the pitching moment coefficient
values by the SST model are below the data till AoA of 20°. Both models overpredict the pitching moment (Cmy)
at 22°. The dip location is predicted by SST around 19° and approximately at 17.5° by the experimental data. The
SA model predicts the dip location around 18.5° with a magnitude smaller than that of experiment. The pitching
moment coefficient is well predicted by SA between 17° and 20° . In Figure 7, the drag moment coefficient
predicted by SST model is good agreement with the experimental data and overpredicted by the SA model. In
general, SST performs better in cp and cd predictions whereas SA does better in Cmy predictions.

Figure 5. Lift coefficient vs. AoA. comparison between CFD (SA, SST) and Experiment

Figure 6. Pitching moment coefficient vs. AoA. comparison between CFD (SA, SST) and
Experiment

5
 Figure 7. Drag coefficient vs. AoA. comparison between CFD (SA, SST) and Experiment

Computed cp distributions for the 53° delta wing SACCON configuration are shown in Figures 8 to 23. Figures 8
to 15 compare the global pressure distributions on the upper wing surface, as predicted by the two turbulence models,
at AoA ranges from 15° to 22°. Corresponding cp profiles at several stations (see Fig. 4) are depicted in Figures 16
to 23. Below AoA = 17° a vortex system starts developing along the leading edge; above AoA= 19° , both
turbulence models predict a strong vortex moving downstream from the leading edge.

Figures 16 to 23 include the PIV measurement to compare against CFD predictions. Both CFD and experiment
predictions include the Cp values of upper and lower surface of the wing. These include span-wise surface pressure
coefficient (Cp) as well as from cross-flow planes (located at x/cr = 0.2, 0.45, and 0.7)

Generally, the SA turbulence model predictions are close to the experimental Cp data than those of the SST
turbulence closure. Therefore only SA model results will be explained in this section. At AoA =15°, there is a good
agreement between the experimental data and the computational predictions.

At AoA=17°, the vortex core strength is over-predicted, thus at x/cr = 0.45 the suction level in the outer vortex
region is under predicted compared to experiment. At AoA=19°, the vortex core is predicted further outboard than
the experimentally observed location. At x/cr = 0.2, 0.45, 0.7 and y/s=0.62, the numerical simulations show that the
primary vortex location shifted to the outboard edge of the wing. Figure 22 shows the pressure distribution at AoA
=22°. As seen in Figure 6, the numerical simulation indicates that pitching moment is over-predicted compared to
the experiment. This corresponds to pressure distributions at x/cr = 0.45 and y/s=0.62 in Fig. 22. Overpredictions of
the suction peak in the outer vortex region contributes an additional nose-down pitching moment.

Figure 8. Pressure contours with SA, a= 15° Figure 9. Pressure contours with SST, a= 15°

6
 Figure 10. Pressure contours with SA, a= 17° Figure 11. Pressure contours with SST, a=17°

Figure 13. Pressure contours with SST, a=19°

Figure 12. Pressure contours with SA, a= 19°

Figure 15. Pressure contours with SST, a= 22°

Figure 14. Pressure contours with SA, a= 22°

7
Figure 16: Surface pressure distribution at 𝜶 =15°, Figure 17: Surface pressure distribution at 𝜶 =15°,
M=0.15, Re=1.6×106. Comparison between CFD M=0.15, Re=1.6×106. Comparison between CFD
(SA model) and Experiment (SST model) and Experiment

8
Figure 18: Surface pressure distribution at 𝜶 =17°, Figure 19: Surface pressure distribution at 𝜶 =17°,
M=0.15, Re=1.6×106. Comparison between CFD M=0.15, Re=1.6×106. Comparison between CFD
(SA model) and Experiment (SST model) and Experiment

9
Figure 20: Surface pressure distribution at 𝜶 =19°, Figure 21: Surface pressure distribution at 𝜶=19°,
M=0.15, Re=1.6×106. Comparison between CFD M=0.15, Re=1.6×106. Comparison between CFD
(SA model) and Experiment (SST model) and Experiment

10
Figure 22: Surface pressure distribution at 𝜶 =22°, Figure 23: Surface pressure distribution at 𝜶 =22°,
M=0.15, Re=1.6×106. Comparison between CFD M=0.15, Re=1.6×106. Comparison between CFD
(SA model) and Experiment (SST model) and Experiment

11
 Figures 24 to 27 show vorticity distributions at constant x sections of the flow-field over the upper side of the
configuration at several angles-of-attack. The seven cut planes located at 0.45, 0.5125, 0.575, 0.6375, 0.7, 0.775, and
0.85 root chord. CFD results are compared with experimental data from PIV measurements. Negative values
indicate that the vortex rotates in clockwise direction. Figure 24 shows that predicted vortex structures agree well
with experimental data. At the higher AoAs, the SST model predicts the vertical structure closer to the data than the
SA model does. This explains the improved lift and drag predictions by the SST closure.

Figure 24. PIV vorticity distribution normal to wing surface, a= 15°

(Top: Experiment, Middle: SA model, Bottom: SST model)

12
Figure 25. PIV vorticity distribution normal to wing surface, a=17°
(Top: Experiment, Middle: SA model, Bottom: SST model

13
Figure 26. PIV vorticity distribution normal to wing surface, a= 19°
(Top: Experiment, Middle: SA model, Bottom: SST model)

14
`

Figure 27. PIV vorticity distribution normal to wing surface, a= 20°

(Top: Experiment, Middle: SA model, Bottom: SST model)

15
 V. Summary and Conclusions

A computational analysis was carried out on the 53° delta wing SACCON geometry to study the effects of
turbulence modeling as well as influence of wing profile on pressure distribution and vortex formation along the
leading edge.

From the numerical computations, conducted using CFD++ with the SA and kω-SST turbulence models, and
comparisons with experimental data, it was concluded that the SA model better predicts moment coefficients
whereas the SST closure predicts lift and drag better. In additions, the SA model predictions are closer to the
experimental cp data than those of the SST closure, yet the latter predicts vortical structures better than the former.

There are two main factors determining these predictions on the delta wing; over-prediction of suction peaks and
incorrect prediction of their location-especially at the higher angles. In a follow on study, grid refinement through
adaptive grid refinement will be utilized. That should enable better prediction of lift, pitching moment, and drag.

Acknowledgments

The authors would like to thank DLR and ONERA teams for providing us wind tunnel test and PIV data.

References
1. LeRoy, J. F., Rodriguez, O. and Kurun, S., “Experimental and CFD contribution to delta wing vortical flow
understanding,” AIAA Paper 2008-0380, 2008.
2. Kruger, W., Hoffmann, D., “Design Consideration for a UCAV Wing for Subsonic and Transonic Aeroelastic
and Flight Mechanic Wind Tunnel Tests,” UAV Design Process/Design Criteria for structure, pp2.3.1-2.3.10.

3. Schütte, A., Cummings, R., Loeser T. and Vicroy, D., “Integrated computational/Experimental Approach to
UCAV and Delta-Canard Configuration Regarding Stability & Control,” 4th Symposium of Integrating CFD
and Experiments in Aerodynamics, Sep. 2009.
4. Nangia, R. K., “Semi-empirical prediction of vortex onset and progression on 65° delta wings (RTO-AVT-
113, VFE-2 facet),” AIAA Paper 2008-0384, 2008.
5. Bennmeddour, Y., Mebarki, Y., Huang X., “Computational Investigation of the Centerbody Effects on the
Aerodynamics of Delta Wing.” Institude for Aerospace research, NRC.
6. Younis, Y., Bibi, A., Khushnood, S.: Vortical Flow Topology on Windward and Leeward side of Delta Wing
at Supersonic Speed,”

7. Schütte, A.; Hummel, D.; Hitzel, S, “Numerical and Experimental Analyses of the vertical flow around the
SACCON configuration,” AIAA Paper 2010

8. Ghaffari, F., “Turbulent Vortex-Flow Simulation over a 65° Sharp and Blunt Leading-Edge Delta wing at
Subsonic Speeds.” NASA/TM-2005-213781.
9. Hitzel, S., “Saccon:Stability Control Configuration Aerodynamics Layout for Test Wing,”

10. Loeser, T., Vicroy, D., Schuette, A., “SACCON Static Wind Tunnel Tests at DNW- NWB and 14´x22´
NASA LaRC,” AIAA Paper 2010-4393, 2010.

11. Gilliot, A., “Static and Dynamic SACCON PIV Tests - Part I: Forward Flowfield,” AIAA Paper 2010-4396,
2010.

16
 12. Konrath, R., Roosenboom, E., Schröder, A., Pallek, D., Otter, D., “Static and Dynamic SACCON PIV Tests
Part II: Aft Flow Field,” AIAA Paper 2010-4396, 2010.

13. Spalart P. R. and Allmaras, S. R., "A One-Equation Turbulence Model for Aerodynamic Flows," La
Recherche Aerospatiale, VOl. 1, 1994.

14. Menter, F. R., "Two-Equation Eddy-Viscosity Turbulence Models for Engineering Applications", AIAA
Journal, vol. 32, pp. 269-289 1994.

15. Chakaravathy, S., “ A unified-grid finite volume formation for computational fluid dynamics., Int. J. Numer.
Meth. Fluids 31:309-323.

16. Batten, P., Leschziner M., Goldberg U., “Average-state Jacobian and implicit methods for compressible
viscous and turbulent flows.”, Journal of Computational Physics 137:38-78.

17. Chakravary S., Peroomian O., Sekar, B., “Some Internal Flow Applications of a Unified-grid CFD
methodology”, AIAA Paper 96-2926.

18. Peroomian. O., Chakravarthy. S., Goldberg. U., “A grid-transparent methodology for CFD,” AIAA Paper 97-
0724.

19. Peroomian O., Chakravarthy. S., Palaniswamy S., Goldberg. U., “Convergence Acceleration for Unified-Grid
Formulation using Preconditioned Implicit Relation,” AIAA Paper 98-0116.

17

View publication stats