UAV SILM logoUAV SILM
CFD

OpenFOAM CFD of the full aircraft

We ran a 3D RANS simulation of the complete V4 airframe at 20 m/s, first with the original wing position and then with the wing moved 20 mm aft. Lift and stability agree closely with XFLR5, and drag is higher, as expected once the fuselage and fully turbulent flow are included. Moving the wing left lift and drag unchanged and moved the neutral point 18.2 mm aft.

0.523
CL at 3° (XFLR5: 0.513)
14.2
L/D at 3°, full aircraft
381.8 mm
neutral point, wing moved aft
20%
static margin at the 343 mm CG
UAV SILM OpenFOAM CFD: surface pressure coefficient on the full aircraft at 3 degrees angle of attack
Drag to rotate

OpenFOAM results for the wing-aft geometry at 20 m/s. The particles move at about 1/16 of real speed and follow the computed streamlines. The aircraft pitches to the chosen angle of attack so the oncoming air stays horizontal. Blue on the surface is suction, red is stagnation.

After the balance check we moved the wing 20 mm aft in CAD and ran the same CFD again at 0° and 3°.

Original wingWing 20 mm aft
CL at 0° / 3°0.245 / 0.5250.243 / 0.523
CD at 0° / 3°0.0275 / 0.03670.0276 / 0.0368
Neutral point363.6 mm381.8 mm
Static margin, same components13.2% (CG 337.8 mm)20.0% (CG 342.7 mm)
Trim speed, ruddervators neutralabout 15 m/s19.4 m/s
Two charts: pitching moment against angle of attack for the original and the moved wing, and trim speed against centre of gravity position
Left: a steeper slope means a more stable aircraft. Right: trim speed falls as the CG moves aft, from 21.2 m/s at 335 mm to 18.8 m/s at 345 mm.

Lift and drag barely changed, so the performance figures still hold. The neutral point moved almost as far as the wing, while the CG moved only about 5 mm. That raised the static margin from 13% to 20% and brought the trim speed close to our 20 m/s cruise target, without adding any weight.

The first run, on 3 November, used the original wing position. Pitching moments are about x = 0.335 m, the CG of our earlier XFLR5 design point.

Angle of attackCLCDL/DCmXFLR5 CL / CD
0°0.2450.02758.9+0.0220.257 / 0.015
3°0.5250.036714.3−0.0190.513 / 0.0227
6°0.7810.056213.9−0.060not run
UAV SILM lift curve, drag polar, lift-to-drag ratio and pitching moment from OpenFOAM compared with XFLR5
OpenFOAM (blue) against XFLR5 (amber). Click to enlarge.
Lift
Matches XFLR5
Hover or tap

Within 5% at 0° and 3°. The lift slope is 0.089 per degree, so the wing and tail sizing from XFLR5 holds up in full 3D flow.

Stability
Stable, but not enough
Hover or tap

The pitching moment falls as angle of attack rises, so the aircraft returns to trim after a disturbance. But the neutral point came out at 363.6 mm, further forward than XFLR5 predicted, so the balance needed more care. See Wing moved aft.

Drag
Higher than XFLR5
Hover or tap

XFLR5 models only the wing and tail with some laminar flow. The CFD adds the fuselage, nose and tail boom and treats the boundary layer as fully turbulent, which overestimates friction. The real value likely lies between the two; flight tests will tell.

UAV SILM OpenFOAM CFD: surface pressure coefficient on the full aircraft at 3 degrees angle of attack, 20 m/s
Surface pressure at 3°. Blue is suction, red is stagnation.
UAV SILM CFD top view surface pressure at 3 degrees
Top view at 3°: suction over the whole upper wing, weakening toward the tips.
Pressure coefficient around the MH32 wing section at mid-span, 3 degrees
Mid-span section at 3°: stagnation point at the nose, suction peak just behind it.
Velocity field around the wing section at mid-span, 3 degrees
Mid-span velocity at 3°: flow stays attached to the trailing edge.
Velocity on the aircraft centreline around the fuselage, 3 degrees
Centreline velocity around the fuselage at 3°.
Streamlines over the UAV SILM wing and fuselage, coloured by velocity, 3 degrees
Streamlines at 3°, coloured by speed.
Wing-tip vortex streamlines on UAV SILM at 6 degrees
At 6°, air from under the tip rolls up into a helical vortex.
Streamwise vorticity in the wake plane behind UAV SILM at 3 degrees
Wake plane at 3°: tip vortices, wing wake and V-tail root vortices.

The tip vortex is where most of the induced drag comes from. It is why drag grows faster than lift at higher angles of attack, and why best L/D occurs around 3 to 4° rather than higher. You can see it in the 3D view above: the particles leaving the wing tips curl around each other.

SoftwareOpenFOAM v2412
SolversimpleFoam, steady RANS
Turbulence modelk-omega SST
Wall treatmentSpalding wall function
Mesh1.27 M cells, snappyHexMesh
Boundary layers3 prism layers
Domainhalf model, symmetry plane
Airspeed / air20 m/s, sea level
Reference area / MAC0.263 m² / 0.195 m
Moment referencex = 0.335 m
Cases0°, 3°, 6°; wing aft 0°, 3°
Neutral point fromslope of Cm against CL
  • Each case ran until residuals had fallen about five orders of magnitude and lift had stopped changing.
  • We ran one mesh resolution, without a mesh-independence study, so we treat drag as accurate to roughly ±15%.
  • The model assumes turbulent flow everywhere, which overstates skin-friction drag at this Reynolds number (about 270,000).
  • The hinge gaps at the aileron and ruddervators are smaller than the mesh cells and are not resolved.
  • For the wing-aft case we moved the wing and spar 20 mm and sealed the wing root with a short extension; the real redesign moves the wing saddle with the wing.

SILM Buteo is in development and has not flown yet. Figures are design estimates from CFD and calculation, not flight-test data. About the student project