Aneesh Boravkeaneesh.boravke.uk

Projects02

SAR UAV: Wing Design & Structural Analysis

Designed and validated the wing for a Snowdonia search-and-rescue UAV, from aerofoil selection to a 1:5 scale model, with no prior tool experience.

The 1:5 scale wing model on a grey backdrop, six laser-cut plywood ribs threaded on a square wooden spar
Fig. 1: My 1:5 scale wing model: six laser-cut plywood ribs on a square wooden spar
Module
FEEG1201
Year
Year 1, 2024–25
Team
5-person team
My part
Wing design and structural analysis
Tools
XFLR5 · Python · laser cutting

1 · Context

5-person team designing a fixed-wing UAV for Snowdonia mountain search and rescue. Brief: a 36-unit fleet, LiDAR + high-res camera + FPV, under 10 kg, high-altitude long-endurance flight over mountainous terrain. The team chose its concept with a weighted decision matrix (aerodynamic efficiency, high aspect ratio, stability, ease of manufacture): Concept 4 scored 33, against 29, 28 and 22, with front and rear contra-rotating propellers, FPV, LiDAR and thermal cameras, and long slender wings. My part was wing design and structural analysis.

2 · What I did

  • Tested 11 NACA aerofoils in XFLR5 at Re = 150,000 (calculated for cruise at 1 km altitude over Snowdonia) and selected NACA 4312 for the highest Cl_max.
  • Designed 15 wing planform iterations to raise Cl/Cd at cruise. The elliptical Design 1 had the lowest, 26.7, and I explored root dihedral for stability. Sweep came out of my XFLR5 testing, new versus the group design.
  • Wrote a custom Python script (trapezoidal rule) for bending moment, slope and deflection along the span.
  • Compared Aluminium 7075-T6 (E = 71.1 GPa) against Carbon Fibre M60J (E = 588 GPa), chose M60J, and designed an I-beam spar cross-section.
  • Built a 1:5 scale physical model. The wing tapered so sharply that the tip parts were too small for the laser cutter and kept breaking, so I doubled the model’s scale to make it buildable. Seeing the real tip come out far smaller than it looked in CAD taught me more than the simulation did.
NACA 4312m = 4% c · p = 30% c · t = 12% cmax camber 4% c at x/c = 0.3max thickness 12% c at x/c ≈ 0.30LETE0.000.250.500.751.00x/cchord linemean camber line
Max camber
4% c at x/c = 0.3
Max thickness
12% c at x/c ≈ 0.30
Solid line
chord, LE to TE
Dashed line
mean camber line
Fig. 2: The selected NACA 4312 aerofoil, drawn from the NACA 4-digit equations: maximum camber 4% of chord at 30% chord, maximum thickness 12% of chord
XFLR5 plot of lift-to-drag ratio against lift coefficient for several planform iterations, each curve peaking between about 32 and 37 near CL 0.5
Fig. 3: XFLR5, Cl/Cd against CL across planform iterations

2.1 · The final wing

Span
3.3 m
Root / tip chord
0.27 m / 0.16 m
Aspect ratio
15.9
Taper ratio
0.37
Root-to-tip sweep
5.8°
Winglet dihedral
40°
Aerofoil
NACA 4312
Spar
Carbon fibre M60J, I-beam
CAD render of the final wing seen from above: tapered, slightly swept, with upturned winglets at the tips
Fig. 4: Final wing from above: taper, 5.8° sweep and winglets

2.2 · The 1:5 model

The model has 6 ribs laser-cut from 3 mm plywood on a 3 mm square wooden spar, because a scaled I-beam was too small to make. I cut the rib holes 0.2 mm wider than tall so the spar could sit at an angle. It shows spar position and rib spacing like an x-ray view.

The 1:5 model from above: plywood ribs spaced along a square wooden spar
Fig. 5: The model from above: rib spacing and spar position
Drawing of the I-beam spar cross-section: 10.00 wide flanges 1.50 thick, 1.50 web, 15.00 overall height
Fig. 6: The full-size I-beam spar section the model’s square spar stands in for

3 · Result

A fully justified wing design with documented structural analysis and a physical model, done end to end in first year.