Supercavitation

How supercavitation works

Water is roughly 800 times denser than air, and the price you pay for moving through it is skin friction — every square centimeter of wetted hull drags. Now imagine erasing nearly all of that wetted area. If you could surround the body in a bubble of gas and touch the water only at a single point on the nose, the friction would almost vanish. That is exactly what supercavitation does.

The trick rests on a simple fact about boiling. Water doesn't only boil when you heat it; it also boils when you drop the pressure. Push a blunt shape through water fast enough and the flow accelerating around its edge drops the local pressure below water's vapor pressure (about 2.3 kPa at 20 °C). The water flashes to vapor, and a cavity forms. At ordinary speeds you get a messy cloud of small bubbles — ordinary cavitation, the kind that eats propellers. Push harder and those bubbles merge into one continuous cavity. Push harder still and that single cavity grows until it sheaths the entire body, closing far downstream in the wake. The body is now flying through gas, wetted only where its nose cavitator pierces the front of the cavity.

The governing dimensionless group is the cavitation number:

       p∞ − pc
  σ = ───────────
        ½ ρ V²

  p∞ = ambient (hydrostatic) pressure      [Pa]
  pc = pressure inside the cavity           [Pa]
  ρ  = water density (≈ 1000 kg/m³)
  V  = vehicle speed                        [m/s]

  Natural supercavitation: pc ≈ p_vapor (~2.3 kPa)
  A body-enveloping cavity needs σ ≲ 0.1.

Read σ as a tug-of-war: the numerator is the pressure trying to crush the cavity shut; the denominator is the dynamic pressure of the flow that wants to hold it open. When σ is small, the flow wins and the cavity stays big. Because the denominator scales with V², σ falls as 1/V² — doubling speed quarters σ. That single fact explains everything downstream: supercavitation is intrinsically a high-speed game, and the only way to cheat it at low speed is to raise pc by pumping gas into the cavity (ventilated supercavitation, below).

The cavitator and cavity geometry

The single most important part is the cavitator — a blunt disk (or shallow cone) at the very tip of the nose. Its sharp edge forces the flow to separate cleanly at a fixed line, pinning the front of the cavity in place. The cavitator is the only part normally touching water, so it carries almost all the drag and, when tilted, generates the lift used to steer. A widely used engineering estimate (Garabedian / Logvinovich) ties the steady cavity size to the cavitator diameter and σ:

Disk-cavitator drag:
  C_d ≈ C_d0 (1 + σ),   with C_d0 ≈ 0.82 for a flat disk
  D   = C_d · ½ ρ V² · A_cavitator

Maximum cavity diameter (Dc) relative to cavitator diameter (dn):
  Dc / dn ≈ sqrt( C_d / (k·σ) )            (k ≈ 1, σ small)

Cavity slenderness (length / diameter) grows as σ shrinks:
  Lc / Dc  ∝  1 / sqrt(σ)

So the cavitator is a sizing knob. A bigger disk makes a bigger cavity (good — it can sheath a fatter body) but also more drag (bad). As σ drops with speed, the same cavitator throws a cavity that is both wider and far more slender — long and cigar-shaped. The designer's job is to keep that cavity just large enough to clear the hull at every operating point, with margin, while spending as little cavitator drag as possible.

Natural vs ventilated supercavitation

There are two ways to get σ below 0.1, and real vehicles use both at different phases of a run.

• Natural (vaporous) supercavitation. The cavity is filled with water vapor at pc ≈ 2.3 kPa. With pc essentially fixed and tiny, the only way to push σ down is to fly fast. At 10 m depth the ambient pressure (atmospheric plus hydrostatic) is about 200 kPa, so σ = 0.1 needs ½ρV² ≈ 2 MPa, i.e. V ≈ 63 m/s (~122 knots). Natural supercavitation is the steady-state regime of a torpedo already at full speed.
• Ventilated (artificial) supercavitation. Inject gas — engine exhaust or stored air — through ports just behind the cavitator. This raises pc, which shrinks the numerator (p∞ − pc) and lets a full cavity form at far lower speed. This is how a vehicle establishes its cavity at launch and holds it during acceleration before natural cavitation can take over. The penalty: you must carry a gas supply, meter it precisely against a ventilation coefficient Cq = Q / (V·dn²), and live with a buoyant, upward-curving cavity that complicates trim.

A subtle gotcha unique to ventilated cavities: at certain gas-injection rates the cavity sheds its gas in periodic pulses (pulsating cavity) rather than closing smoothly (re-entrant jet closure or twin-vortex closure). Pulsation shakes the vehicle and is actively avoided by tuning the ventilation rate and the depth-dependent Froude number.

Worked example: speed needed at depth

Take a torpedo running at 15 m depth and ask how fast it must go for natural supercavitation.

Hydrostatic + atmospheric pressure at 15 m:
  p∞ = p_atm + ρ g h
     = 101.3 kPa + (1000)(9.81)(15)
     = 101.3 kPa + 147.2 kPa = 248.5 kPa

Cavity (vapor) pressure:  pc ≈ 2.3 kPa
Numerator:  p∞ − pc ≈ 246.2 kPa

Target σ = 0.08 (comfortably enveloping cavity):
  ½ ρ V² = (p∞ − pc) / σ = 246.2 kPa / 0.08 = 3.08 MPa
  V² = 2 × 3.08e6 / 1000 = 6.16e3
  V  = 78.5 m/s  ≈ 153 knots

Now the drag payoff. Suppose the body has a frontal area of 0.05 m² (a ~250 mm torpedo). Fully wetted at 78.5 m/s with a turbulent skin-friction-dominated C_d ≈ 0.10 over a much larger wetted area, drag is tens of kilonewtons. Supercavitating, only the cavitator is wetted. With a 40 mm disk cavitator (A ≈ 1.26 × 10⁻³ m²) and C_d ≈ 0.82(1 + 0.08) ≈ 0.89:

D = C_d · ½ ρ V² · A_cav
  = 0.89 × 3.08e6 × 1.26e-3
  ≈ 3.45 kN

A fully-wetted body of the same diameter at the same speed
faces ~10× this — the wetted area, not the nose, dominates.
That order-of-magnitude drag cut is the whole point.

The catch the numbers hide: pushing a 250 mm body at 78 m/s still needs roughly P = D·V ≈ 3.45 kN × 78.5 m/s ≈ 270 kW just against cavitator drag, which is why supercavitating weapons use rocket motors, not propellers — a propeller can't bite the gas inside the cavity.

Real-world examples

System	Type	Speed	Notes
VA-111 Shkval (USSR/Russia)	Rocket torpedo	~200 kn (~100 m/s)	Solid-rocket powered, ventilated then natural cavity; straight-running, short range (~7–13 km)
Hoot / Valfajr (Iran)	Rocket torpedo	~200 kn (claimed)	Reported Shkval-derived design
Barracuda (Germany, Diehl R&D)	Test vehicle	~400 km/h (research)	Demonstrated guided supercavitating maneuvering
Supercavitating projectile (Navy ONR)	Bullet / round	>1,000 m/s entry	RAMICS / ADAPS-type rounds for mine and diver defense
Supercavitating propeller	Marine propeller	40+ kn craft	Blades run with attached cavity on purpose to avoid erosion at high tip speed
High-head hydro / pump research	Test rig	n/a	Ventilated cavities studied to avoid blade erosion in turbines and inducers

Supercavitation vs other underwater drag states

	Fully wetted	Partial cavitation	Supercavitation (natural)	Supercavitation (ventilated)	Surface planing
Wetted area	Whole hull	Hull minus cavity patch	Nose cavitator only	Nose + tail planing patch	Hull bottom only
Cavitation number σ	> ~1	~0.3 to 1	≲ 0.1	≲ 0.1 (via gas)	n/a (free surface)
Cavity contents	None	Vapor, unsteady	Water vapor	Injected gas/air	Air above water
Dominant drag	Skin friction	Friction + pressure	Cavitator pressure drag	Cavitator + ventilation	Wave + spray
Speed regime	Any	Moderate–high	Very high (>~50 m/s)	Low–high (gas-assisted)	High (planing craft)
Erosion risk	None	Severe (bubble collapse)	None (cavity closes in wake)	None	Spray wear only
Steering authority	Full (fins work)	Reduced near cavity	Cavitator + piercing fins only	Cavitator + planing forces	Rudder / strakes
Typical home	Submarines, ships	Loaded propellers, pumps	Rocket torpedoes, projectiles	Launch/accel phase, test vehicles	Hydrofoils, race boats

Stability, planing, and control

Inside the cavity the vehicle is surrounded by gas, so almost nothing on the body produces hydrodynamic force. Conventional fins are useless unless they pierce the cavity wall into the surrounding water. The control authority comes from just two places: the nose cavitator (deflect it and the whole nose generates lift) and tail surfaces designed to poke through the cavity boundary. This is a fundamentally underactuated, nonlinear control problem.

Worse, the vehicle planes. Because the cavity is slightly buoyant and the body is denser than gas, the tail tends to fall onto the cavity's lower water interface, slap it, get pushed back up, and slap again. This tail-slap creates a strong, history-dependent restoring force that is highly nonlinear — small changes in immersion depth cause large jumps in force. Control engineers model it as a switched system and design controllers (often gain-scheduled or sliding-mode) that modulate cavitator angle and gas-injection rate to hold a stable limit cycle rather than fight the planing entirely. Get it wrong and the vehicle pitches hard, the cavity collapses asymmetrically, and the body is suddenly slammed by full wetted drag — a failure that can break the vehicle in milliseconds.

When supercavitation is worth it

• When raw underwater speed dominates all other goals — interception weapons, mine-clearance rounds, and short-range torpedoes where a few seconds of warning matter more than range, quietness, or guidance.
• When the propulsion can run without water — rocket motors and gas-generator drives work inside the cavity; conventional propellers cannot, because they would spin in gas.
• When you can tolerate short range — cavitator drag still costs hundreds of kilowatts, and rockets burn out fast, so supercavitating weapons trade range for speed (Shkval ≈ 7–13 km vs 30+ km for slow torpedoes).
• When erosion at high speed is otherwise unavoidable — supercavitating propellers deliberately run with an attached cavity so the bubbles close in the wake instead of pitting the blade, which is the high-tip-speed alternative to fighting partial cavitation.

Avoid it when you need stealth (rockets and gas venting are loud and leave a wake), long range, precise terminal guidance, or low speed. A quiet, fully-wetted, propeller-driven torpedo still wins almost every mission except the dash.

Common misconceptions and pitfalls

• "Supercavitation breaks the underwater sound barrier." No. The often-quoted "supersonic" tag refers to research projectiles whose entry speed can exceed the ~1,500 m/s speed of sound in water for a brief, decelerating dash — it is not a steady cruise condition, and a 200-knot torpedo is nowhere near it. The real win is drag, not a sonic threshold.
• "The bubble is air the vehicle carries." In natural supercavitation the cavity is water vapor created by the flow itself, not stored air. Only ventilated supercavitation adds gas, and even then the cavity is mostly maintained by the flow geometry, not by carrying a giant air tank.
• "Less drag means it's easy to go fast." The drag cut is real, but cavitator pressure drag is stubborn and rises with σ, and you still need hundreds of kilowatts. The hard limits are propulsion energy density and control, not friction.
• "You can just add fins to steer it." Fins inside the gas cavity do nothing. Only the cavitator and cavity-piercing surfaces have authority, which is why these vehicles historically ran nearly straight and only recently gained limited maneuvering.
• "Cavitation and supercavitation are the same problem." Ordinary cavitation collapses bubbles on the surface and erodes it; supercavitation deliberately grows one cavity that closes downstream, so it avoids erosion entirely. They are opposite design intents using the same physics.
• "Going deeper helps." The opposite. Depth raises p∞, which raises σ at fixed speed, shrinking the cavity. Supercavitating weapons run shallow; the same vehicle that supercavitates at 10 m may not at 100 m without far more speed or ventilation.