Vortex Generators

How a vortex generator works

Air flowing over a wing doesn't slide along the surface as one block. Right at the skin, friction drags the air to a near-standstill, and over a thin region — the boundary layer — the speed climbs from zero at the wall to the full free-stream value a few millimeters or centimeters out. That near-wall air is sluggish. It has very little momentum, and it's the first thing to give up when the wing asks it to do something hard.

The hard thing is the rear of the wing. Past the point of maximum thickness, the surface curves back inward and the static pressure rises along the flow direction — an adverse pressure gradient. The boundary layer now has to flow "uphill" against rising pressure. The fast outer air manages it; the slow near-wall air doesn't. When that slow layer runs out of momentum it stops, reverses, and lifts off the surface. That's separation, and when it spreads forward across the wing it's a stall: lift collapses, drag spikes, and the controls go mushy.

A vortex generator fixes this without touching the airfoil shape. It's a small vane — a flat or cambered plate, often the height of the boundary layer itself — bolted or bonded to the surface and set at an angle of attack of roughly 10° to 25° to the local flow. Because it has a sharp edge and an angle, it sheds a streamwise vortex, a tight tube of swirling air trailing downstream like a miniature tornado lying along the surface.

That vortex is a mixing pump. As it spins, it continuously scoops high-momentum air from the outer flow and drives it down to the wall on one side, while lifting the tired near-wall air up and away on the other. The boundary layer downstream of the VG is no longer a clean stack of slow-over-fast layers — it's been stirred, and the near-wall air now carries far more momentum than it did. With that extra momentum it can climb the adverse pressure gradient that would otherwise have stopped it. Separation is pushed aft, or delayed to a higher angle of attack, or eliminated entirely in the region the array covers.

The key insight: a VG adds no energy to the flow. It only redistributes momentum that was already there, taking it from the abundant outer flow and delivering it where it's scarce. That's why it's so cheap — a passive vane a centimeter tall — and also why it isn't free: stirring the flow costs a little drag whether the wing needed it or not.

The governing physics

The whole story lives in the momentum of the near-wall flow versus the pressure it has to climb. Separation occurs where the wall shear stress drops to zero — where the velocity profile's slope at the wall goes flat:

Wall shear stress:   τ_w = μ · (∂u/∂y)|_(y=0)
Separation point:    τ_w = 0   ⇒   (∂u/∂y)|_(y=0) = 0   (incipient reverse flow)

Adverse pressure gradient drives separation:  dp/dx > 0
Boundary-layer momentum (von Kármán integral):
    d θ/dx + (2 + H)·(θ/U)·(dU/dx) = C_f / 2
      θ = momentum thickness,  H = shape factor (δ*/θ),  C_f = skin-friction coeff.

Separation is imminent as the shape factor H rises:
    H ≈ 1.3 to 1.4 (turbulent attached)  →  H ≈ 2.2 to 2.4+ (turbulent separating)

A VG works by lowering the shape factor H downstream: by injecting momentum near the wall it makes the velocity profile fuller (more like a flat-top), which keeps ∂u/∂y at the wall positive further into the adverse gradient. The vortex strength scales with the vane's circulation:

Vane circulation (lifting-line / Kutta–Joukowski estimate):  Γ ≈ ½ · C_L,vane · U · c
    c        = vane chord (streamwise length)
    h        = vane height (≈ boundary-layer thickness δ)
    C_L,vane = vane lift coefficient at its incidence (~0.5 to 1.0)
    U        = local edge velocity

Boundary-layer thickness (turbulent, flat plate):
    δ ≈ 0.37 · x · Re_x^(-1/5),   Re_x = U·x/ν

So the design height isn't arbitrary: it's tied to δ, which itself grows with distance x back from the leading edge and shrinks at higher Reynolds number. A vane much shorter than δ sits buried in slow air and sheds a feeble vortex; a vane much taller than δ pokes into flow that didn't need mixing and just makes drag. The sweet spot — repeatedly confirmed in wind-tunnel studies (Lin, NASA Langley, 1990s–2000s) — is a vane height around 0.8 to 1.2 δ, spaced roughly 4 to 6 vane-heights apart along the span, placed just upstream of the natural separation line.

Design tradeoffs and failure modes

• Always-on drag. A VG array adds profile drag every second of flight, including cruise where the flow was happily attached and needed no help. A typical wing-array penalty is a small fraction of a percent of total drag, but on a long-haul airliner that compounds into measurable fuel burn. Designers minimize it by sizing vanes as small as the job allows — hence the interest in low-profile / sub-boundary-layer VGs at 0.1 to 0.5 δ that cut the penalty roughly in half while keeping most of the separation control.
• Placement is everything. Put the array too far aft and the boundary layer has already separated before the vortex can help; too far forward and the vortex diffuses and loses authority before it reaches the trouble spot. The window is usually just upstream (a few percent of chord) of the predicted separation line. Get it wrong and the VGs add drag for no benefit.
• Sizing for the wrong condition. A vane sized to the cruise boundary layer is too short for the thicker, lower-Reynolds-number boundary layer at approach speed — exactly when you most need stall margin. VGs are usually sized for the critical low-speed, high-angle-of-attack case.
• Direction sensitivity. Counter-rotating pairs are tuned for one flow direction; if the local flow angle swings (near a deploying flap, a sweeping wing root, or a yawing fin) the vortex pattern degrades. Co-rotating arrays tolerate flow-angle change better, which is why they show up where direction varies.
• Contamination and damage. A bent, missing, or ice-clogged vane is a dead spot in the array. Stick-on VGs can peel; metal ones can fatigue at the root. Because the array's job is safety margin near stall, missing vanes are tracked on inspection.
• Erosion. On rotor blades and wind-turbine blades, the leading-edge region erodes from rain and grit, and VGs in that zone wear or detach. Replaceable bonded strips are common so the array can be refreshed without a blade rebuild.

Types of vortex generator

	Counter-rotating vanes	Co-rotating vanes	Low-profile (sub-δ)	Wheeler / wishbone (doublet)	Air-jet VGs
Vortex pattern	Mirrored pairs, alternating up/down-wash	All same sense, spanwise drift	Same as vane type but weaker	Embedded ramp pair, low height	Vortex from a blown jet, no vane
Typical height	~1.0 δ	~1.0 δ	0.1 to 0.5 δ	0.2 to 0.5 δ	None (flush hole)
Drag penalty	Moderate	Moderate	Low	Low	Zero when off
Mixing per unit drag	Highest (fixed direction)	Good, direction-tolerant	Lower authority	Good for the height	Tunable, can switch off
Flow-angle tolerance	Low	High	Moderate	Moderate	High
Complexity	Passive plate	Passive plate	Passive plate	Passive ramp	Needs bleed air + plumbing
Typical home	Airliner wings, tails	Diffusers, varying-flow ducts	Sailplanes, drag-critical wings	Wind tunnels, research	Engine inlets, active control

Real numbers and specifications

Some figures that anchor the device in reality:

Parameter	Typical value	Notes
Vane height	~0.8 to 1.2 × δ	≈ 1 to 5 cm on transport jets; 2 to 10 mm on sailplanes, turbine blades
Vane incidence to local flow	10° to 25°	Higher angle = stronger vortex but more drag and risk of the vane itself stalling
Spanwise spacing	~4 to 6 vane heights	Close enough for adjacent vortices to maintain a continuous re-energized layer
Chordwise placement	Just upstream of separation line	Often 10% to 60% chord depending on airfoil and condition
Stall-angle increase	~2° to 4°	Pushes CL,max up; lowers stall and approach speed by a few knots
Cruise drag penalty	Order of 0.1% to a few % of total drag	Depends on array size; low-profile VGs roughly halve it
Aspect ratio of vane	Height-to-length ≈ 1:2 to 1:4	Long-enough chord to develop the design circulation Γ

Where vortex generators are used

• Transport-jet wings and tails. Spanwise rows of small vanes upstream of the ailerons and over the outer wing keep the flow attached at high lift and improve roll authority near stall. They're also common ahead of the horizontal tail to prevent tail-buffet at high angle of attack. Many production airliners and business jets carry factory VGs.
• STOL and bush-plane retrofits. After-market stick-on VG kits (Micro AeroDynamics, Stene, and others) are a popular mod on light aircraft and bush planes. By lowering stall speed a few knots they shorten takeoff and landing rolls and tame low-speed handling — a cheap upgrade compared with new flaps or a new wing.
• Sailplanes and gliders. Tiny low-profile VGs on the wing improve climb and low-speed handling and can cure a too-early root separation that hurts thermalling. Because cruise (high-speed glide) drag matters intensely, glider VGs are kept as small as possible.
• Wind-turbine blades. VG strips bonded to the inboard, thick part of the blade keep flow attached there, raising power capture in the partial-load region and reducing load fluctuations. Retrofitting VGs is one of the cheapest blade-performance upgrades and is widely done in service.
• Helicopter and propeller blades. VGs delay retreating-blade or root-region separation, smoothing the rotor's behavior near its lift limit.
• Engine inlets and diffusers. Inside an S-duct inlet or a wide-angle diffuser, the flow wants to separate off the inner wall; VGs (vane or air-jet type) keep it attached so the engine sees clean, uniform flow. Air-jet VGs are attractive here because they can be switched off when not needed.
• Cars and trucks. Small roof-edge VGs (famously on some hatchbacks and the Mitsubishi Lancer Evolution) keep the flow attached over the rear glass to reduce wake drag and keep the rear window cleaner; trailer-top VGs do the same for trucks.

Vortex generators vs other separation-control devices

	Vortex generators	Boundary-layer suction	Blown / Coandă flap	Leading-edge slat	Riblets
What it controls	Separation (re-energizes BL)	Separation (removes slow air)	Separation + adds lift	Separation at high α	Skin-friction drag, not separation
Adds energy to flow?	No (redistributes)	Yes (pump work)	Yes (engine bleed)	No (re-shapes flow)	No
Active or passive	Passive (vanes)	Active	Active	Passive/mechanical	Passive
Complexity	Lowest	High (pumps, ducts)	High (plumbing, control)	Moderate (tracks, actuators)	Low (surface texture)
Always-on cost	Small parasitic drag	Pump power	Bleed-air thrust loss	Stowed in cruise	Tiny
Stall-angle gain	~2 to 4°	Several degrees	Large (high-lift)	Large	None
Typical home	Wings, blades, ducts	Research, some laminar wings	STOL, naval aircraft	Airliner leading edges	Some aircraft, ship hulls

Common misconceptions and pitfalls

• "VGs add energy to the boundary layer." They don't. They redistribute momentum already present in the flow, pulling it down from the outer stream. Saying they "energize" the boundary layer is shorthand — there's no power input from a passive vane.
• "Bigger vanes work better." Only up to about the boundary-layer height. Beyond that you add drag without adding useful mixing, because the extra vane height sits in flow that was never going to separate. Many real installs trend smaller, not larger, to chase the drag penalty down.
• "VGs increase maximum lift by adding lift themselves." The vanes' own lift is negligible. The lift gain comes from keeping the wing's flow attached to a higher angle of attack, so the wing reaches a higher C_L,max before it stalls. The VG is an enabler, not a lifting surface.
• "They're the same as a winglet." No — a winglet fights the wing-tip vortex to cut induced drag in cruise; a VG deliberately makes a small streamwise vortex to fight chordwise separation. Different vortices, different jobs.
• "More VGs is always safer." Past the optimum spacing the array stops improving and just costs drag. And an over-mixed boundary layer is a thicker, draggier one. The right answer is the smallest array, in the right place, that cures the separation.
• "VGs only matter at stall." They also fix local separation that hurts in normal flight — tail buffet, an aileron with poor authority, an engine inlet feeding distorted flow to the fan, a diffuser losing pressure recovery. Stall delay is the headline use, not the only one.