Winglet

How a winglet works

Every wing that makes lift pays a tax at its tips. The pressure under the wing is higher than the pressure on top — that pressure difference is the lift — and at the open end of the wing nothing stops the high-pressure air from spilling around the tip into the low-pressure region above. That spill rolls up into a tight, spinning tube of air that streams off behind the aircraft: the trailing wingtip vortex. Each vortex flings a wide sheet of air downward (the downwash), which tilts the local airflow — and therefore the lift vector — slightly backward. The rearward component of that tilted lift is induced drag, and it can be a third or more of the total drag in climb.

A winglet is a small, carefully shaped lifting surface stood up at the tip, sitting right in the swirling cross-flow of the vortex. It does two things at once:

• It generates a forward force. The vortex's circular flow hits the winglet at an angle, so the winglet produces a side-force whose component points forward along the flight path — a small thrust that directly offsets drag, exactly the way a sailboat beats upwind.
• It moves and spreads the vortex. By carrying the wing's loading further outboard and upward, the winglet pushes the strongest part of the vortex out past the physical tip and diffuses it, so the downwash over the main wing is gentler. Less downwash means the lift vector tilts back less, which is the same as saying induced drag falls.

In the language of wing efficiency, a winglet raises the span efficiency factor e (and the effective aspect ratio), which is the single lever in the induced-drag equation. It is not magic and it is not a thrust generator in the propulsive sense — it simply recovers energy that the wing was throwing away into the vortex.

The governing math

Induced drag is captured by the classic lifting-line result. For a wing of lift coefficient C_L and aspect ratio AR:

Induced-drag coefficient:

  C_Di = C_L² / (π · AR · e)

  AR = b² / S        (b = span, S = wing area)
  e  = span efficiency factor  (≈ 0.75–0.85 typical airliner; 1.0 = ideal elliptical)

Induced drag force:

  D_i = ½ · ρ · V² · S · C_Di
      = (2 · L²) / (π · ρ · V² · b² · e)

Read those two relations carefully and the whole winglet story falls out:

• Induced drag scales with the square of lift (or weight) and inversely with the square of span and with V². So it is worst when the aircraft is heavy and slow — climb-out — and shrinks as fuel burns off and speed builds in cruise.
• A winglet doesn't change C_L or the gross span much, so it works through e and the effective AR. Raising e from, say, 0.78 to 0.85 lowers C_Di by about 8% with no change in span.

A useful way to compare a winglet of height h against simply extending the span is the k-factor rule of thumb: a vertical winglet of height h gives an induced-drag benefit roughly equivalent to a horizontal span extension of about k·h, with k ≈ 0.4 to 0.5. So a 2.5 m winglet behaves a bit like adding ~1 m of span per side — but it does it vertically, staying inside the gate box, which is the entire point.

Worked example: a narrow-body in climb

Take a loaded single-aisle airliner just after takeoff and run the numbers for the wing alone:

Weight (lift needed):  L = W = 70,000 kg × 9.81 = 687 kN
Wing area:             S = 124 m²
Span (no winglet):     b = 34.1 m  →  AR = b²/S = 9.4
Air density at climb:  ρ = 0.9 kg/m³   (≈ 10,000 ft)
Speed:                 V = 110 m/s     (≈ 214 kt)

Dynamic pressure:      q = ½ρV² = 0.5 × 0.9 × 110² = 5,445 Pa
Lift coefficient:      C_L = L / (q·S) = 687,000 / (5,445 × 124) = 1.02

Baseline span eff.:    e = 0.78
  C_Di = C_L² / (π·AR·e) = 1.02² / (π × 9.4 × 0.78) = 0.0451
  D_i  = q·S·C_Di = 5,445 × 124 × 0.0451 = 30.4 kN

With blended winglet:  e rises to ≈ 0.85
  C_Di = 1.02² / (π × 9.4 × 0.85) = 0.0414
  D_i  = 5,445 × 124 × 0.0414 = 27.9 kN

Induced-drag reduction in climb:  (30.4 − 27.9) / 30.4 ≈ 8%

An 8% cut in induced drag during climb translates to roughly a 3 to 5% cut in total drag for that phase, because induced drag is only part of the total. Integrated over a long flight — where climb and early cruise dominate the high-lift portion — that lands the typical 3 to 5% block-fuel saving manufacturers quote for blended winglets. On a 737-800 burning around 2,500 kg of fuel per flight hour, a 4% saving is ~100 kg/hr, or on the order of 100,000+ kg of jet fuel per aircraft per year.

Winglet variants and what they trade

Type	Geometry	Typical fuel saving	Best for	Real examples
Wingtip fence	Small surfaces above and below the tip	~1 to 2%	Minimal span/weight growth, gate-tight	Early A320, A380
Blended winglet	Smooth upward curve to a near-vertical tip	~3 to 5%	Retrofit, span-limited fleets	737NG, 757, 767, BBJ
Sharklet	Blended winglet, factory-integrated	~3.5 to 4%	New-build narrow-body	A320 family
Split scimitar	Blended tip plus a downward scimitar ventral fin	~5 to 6%	Squeezing more from existing blended wing	737-800/-900 retrofit
Raked wingtip	Highly swept, mostly horizontal extension	~4 to 6%	Adding effective AR when span allows	767-400ER, 777-300ER, 787
Spiroid / closed loop	Looped tip joining upper and lower surfaces	research / experimental	Maximum vortex suppression (heavy)	Aviation Partners test fleet

The pattern across the table: the more aggressively a device chases the vortex, the more drag it kills — but the more wetted area, weight, and tip bending/torsion it adds. The split-scimitar exists precisely because a second, downward surface catches the part of the vortex a single upper winglet misses, at the cost of more structure.

Real-world specifications

Aircraft	Device	Approx. height / span add	Quoted benefit	Notes
Boeing 737-800 (NG)	Blended winglet	~2.4 m tall	~4% fuel; 150+ mi (~240 km) extra range	Aviation Partners retrofit, later factory option
Boeing 737-800	Split scimitar	+ ventral fin	~1.5% on top of blended (~5.5% total)	Retrofit onto existing blended winglets
Airbus A320neo	Sharklet	~2.4 m tall	~3.5 to 4% fuel	Span 35.8 m — deliberately under 36 m Code C
Boeing 787-9	Raked wingtip	~2.7 m raked extension	Built into 60 m span design	No vertical winglet; raked tip + flexible wing
Boeing 777-300ER	Raked wingtip	~2 m of raked span per side	Improved takeoff & climb	Replaced earlier 777 tall winglet studies
Airbus A330-200/300	Wingtip fence → Sharklet (A330neo)	fence small; Sharklet ~ taller	fence ~1%; neo Sharklet more	Shows generational shift fence → blended

Note the recurring 36 m span ceiling: the A320 family sits at 35.8 m and the 737 at 35.8 m precisely so they fit an ICAO Code C / FAA Group III gate and taxiway. That constraint is the reason these aircraft grow up with winglets rather than out with longer wings. Larger aircraft like the 777-9 instead use folding wingtips so a 71.8 m span can fold to under 65 m at the gate.

When a winglet pays off

• Long stage lengths. The fuel saving compounds with hours flown, so winglets pay back fastest on aircraft cruising for many hours. A regional jet doing 30-minute hops sees a smaller benefit and may not justify the weight.
• Span-constrained airframes. When the wing already touches the gate-box span limit, a winglet is the only way to buy more effective span. This is the core case for narrow-body retrofits.
• High-lift-coefficient operation. Aircraft that climb heavy, operate from hot-and-high airports, or fly at high weights spend more time where induced drag dominates — exactly where a winglet helps most.
• Retrofit economics. A blended-winglet retrofit on a 737/757 costs roughly US$500k to $1M per aircraft installed; at multi-thousand-dollar daily fuel bills, payback is often 2 to 4 years, after which it is pure margin (plus a marketing "greener" halo).

Skip the winglet when the wing has plenty of span headroom (extend the span instead — it is more efficient per metre), when the aircraft flies very short legs at light weight, or when the structural reinforcement to handle tip loads costs more weight than the drag saving recovers.

Misconceptions and pitfalls

• "Winglets produce thrust." Misleading. They produce a force with a forward component in the vortex flow, which offsets drag — but the net effect is reduced drag, not added propulsion. The engine still does all the pushing.
• "Bigger winglet is always better." No. Beyond an optimum height the added skin-friction and profile drag, plus the structural weight and tip bending/torsion, overtake the induced-drag saving. Winglets are sized for a specific design lift coefficient and cruise condition.
• "Winglets help at all speeds." They help most at low speed / high C_L (climb). At high-speed, low-C_L cruise the induced-drag share is small, and a poorly matched winglet can add net drag through its own wetted area and wave-drag effects near the tip.
• Flutter and load margins. Adding mass and area outboard shifts the wing's mass and aerodynamic centres and can lower the flutter speed. A retrofit usually requires structural reinforcement of the outer wing and recertification — a real cost that has killed several retrofit business cases.
• "They eliminate the vortex." They weaken and displace it, raising span efficiency toward (never reaching) the ideal elliptical value of e = 1. The vortex — and wake turbulence separation behind the aircraft — still exists.
• Confusing winglets with aspect ratio. A winglet adds effective aspect ratio at constant gross span; it is not a substitute for the underlying fact that a genuinely higher-aspect-ratio wing (a glider's) is fundamentally more efficient. Winglets are what you reach for when you cannot have the long wing.