Supercavitation: Wrapping an Object in Its Own Vapor Bubble

From speed to a boiling cavity: the cavitation number

Push a solid through water and the fluid must accelerate to flow around it. By Bernoulli's principle, where the water speeds up its static pressure falls. On the sides and trailing surfaces of a fast body that pressure can drop all the way to the liquid's saturation (vapor) pressure — about 2.3 kPa for water at 20 °C, roughly 1/44 of atmospheric. At that point the water crosses the liquid–vapor phase boundary and “boils” at room temperature, not because it is hot but because it is depressurized. Cold vapor pockets nucleate on the dissolved-gas and micro-bubble nuclei that ordinary water always carries; the latent heat is drawn from the surrounding liquid, which cools very slightly.

Whether this happens is set by a single dimensionless group, the cavitation number σ = (p_∞ − p_v) / (½ ρ V²), which weighs the pressure available to keep water liquid against the dynamic pressure of the flow. Cavitation begins when the flow's minimum pressure coefficient reaches −σ. As velocity V climbs, σ shrinks: the low-pressure vapor region grows from a scatter of bubbles (σ ≈ 0.5–1), to a patch attached to the surface, and finally — near σ ≲ 0.1, and for a full body-length cavity around σ ~ 0.02–0.05 — into one continuous cavity that engulfs the entire object. That end state is supercavitation.

The cavitator: seeding and shaping the bubble

A supercavity does not start itself cleanly; it is seeded by a cavitator, a deliberately blunt nose — usually a flat disk or shallow cone — mounted at the very tip. The disk plants a front stagnation point and presents a sharp rim from which the flow separates crisply, springing the cavity boundary outward and away from the body behind it. A flat-disk cavitator carries a drag coefficient of C_D0 ≈ 0.82 as σ → 0, so its pressure drag is F ≈ ½ρV² · C_D0(1+σ) · (¼πd_n²), where d_n is the disk diameter.

Free-streamline theory then fixes the cavity's size. To leading order the maximum cavity diameter is D_c ≈ d_n√(C_D/σ), while its length grows faster still — roughly as 1/σ with a slow logarithmic correction — results traced to the 1950s free-boundary calculations of Garabedian, Birkhoff and Gilbarg and to Hermann Reichardt's wartime experiments. At σ = 0.03 that makes the cavity about five cavitator-diameters wide and tens of diameters long: a slender vapor ellipsoid inside which the vehicle rides. Logvinovich's principle of the independence of cavity-section expansion — each cross-section grows and shrinks as if it were its own 2-D bubble — turns this into a practical recipe for computing cavity shape along a real trajectory.

Two ways to fill it: vaporous and ventilated cavities

Two regimes reach the same end. In natural (vaporous) supercavitation the cavity is filled by water vapor at pressure ≈ p_v, and the only lever on σ is brute speed. Setting σ ≈ 0.03 at the surface (p_∞ ≈ 1 atm) gives V ≈ √(2(p_∞−p_v)/(ρσ)) ≈ 80 m/s — nearly 300 km/h — just to open the cavity, and far more at depth, where p_∞ rises by about one atmosphere every 10 m and drags the threshold speed up with it.

The workaround is ventilated supercavitation: inject gas — air, or a torpedo's own rocket exhaust — into the wake just behind the cavitator. Now the cavity pressure p_c is set by the gas supply rather than by p_v, so σ_c = (p_∞ − p_c)/(½ρV²) can be forced low at a modest speed. The cost is a continuous gas budget — characterized by a ventilation coefficient C_Q = Q/(V d_n²) — and sensitivity to the Froude number Fr = V/√(g d_n): buoyancy tilts the cavity upward, and at the wrong ventilation rate it begins to pulsate. Russia's VA-111 Shkval rocket torpedo (in service 1977) does precisely this — a disk cavitator plus exhaust ventilation — to hold a cavity at roughly 100 m/s.

Why the drag collapses

The entire point of the exercise is drag. A fully wetted torpedo is dominated by skin friction over its whole surface, and water is punishing: at ~800 times the density of air, the dynamic pressure ½ρV² — and with it the drag it generates — is roughly 800 times greater than in flight at the same speed. Wrap the body in vapor and the wetted area collapses to little more than the cavitator disk. Skin friction over the hull essentially disappears; the drag budget shrinks to the cavitator's pressure drag plus small intermittent planing forces at the tail.

The result is underwater speed that is otherwise unreachable. Conventional torpedoes top out near 30–40 m/s; the Shkval reaches roughly 100 m/s (≈200 knots, 370 km/h). Supercavitating projectiles go further still: fired into water at hundreds to over a thousand metres per second, a shaped supercavitating nose keeps them enveloped for the first stretch of their run — the basis of the US Navy's RAMICS anti-mine round and of Nammo/DSG supercavitating small-arms ammunition designed to fly true through water.

Closure, planing and the steering problem

A supercavity is not a tidy balloon. Downstream it has to close, and it does so in one of two ways: a re-entrant jet, in which liquid sprays back into the cavity from the rear stagnation region (typical of vapor cavities), or a twin-vortex closure, in which ventilating gas drains away down two hollow trailing vortices. Both are unsteady, and ventilated cavities in particular can pulsate — shedding and re-forming in cycles.

Because the body barely touches water, it is also hard to steer. Control forces are available only from the cavitator (which can be deflected like a tiny rudder) and from tail planing — the aft body periodically slapping the cavity wall. This tail-slap both supplies a restoring force and buffets the vehicle; the resulting dynamics are strongly nonlinear and close to neutrally stable, which is why weapons like the Shkval have historically run very fast but nearly straight. And if the cavity ever collapses onto the hull instead of closing cleanly behind it, the impact is violent — the very same erosive mechanism that pits ship propellers.

Ordinary cavitation, and how we know

Supercavitation is a single macroscopic cavity, deliberately made and held. Ordinary (inertial) cavitation is its destructive cousin: clouds of tiny vapor bubbles that form in the low-pressure zones of propeller tips, pump inlets and valves, then are swept into higher pressure and collapse. The radial dynamics of a single bubble follow the Rayleigh–Plesset equation; Lord Rayleigh's 1917 analysis — prompted by the Royal Navy's propeller-erosion problem, brought to him via Charles Parsons — gives a collapse time τ ≈ 0.915 R₀√(ρ/Δp) and pressures reaching gigapascals, with re-entrant microjets near 100 m/s and hot spots of thousands of kelvin that can even flash visible light (sonoluminescence). Same physics — local pressure below p_v — but the opposite intent: here the collapse is the enemy, not the goal.

Both are studied in cavitation water tunnels, closed-loop channels (Parsons built the first in 1895) whose pressure can be lowered independently of flow speed to dial in σ, with high-speed and stroboscopic photography, particle-image velocimetry and hydrophones recording cavity shape, closure and pulsation. The modern theory of free-streamline and supercavitating flows was built by Marshall Tulin (slender-body cavity theory, US Navy, 1950s–60s) and, in parallel, by G. V. Logvinovich and the Soviet school that produced the Shkval. More recent programs — Penn State's Applied Research Laboratory and DARPA's Underwater Express — push toward controllable, vehicle-scale supercavitating craft, where the open questions are precisely the hard ones above: stable ventilation, cavity control, and steering a body that hardly touches the sea.

Ordinary (inertial) cavitation vs. supercavitation — same physics, opposite engineering intent.

Feature	Ordinary cavitation	Supercavitation
Vapor scale	Clouds of microscopic bubbles	One body-length cavity
Cavitation number σ	Near inception (~0.3–1)	Very low (σ ≲ 0.1)
Lifetime	Form and collapse in milliseconds	Stable, continuously sustained
Effect on the body	Erosion, noise, loss of lift/thrust	Drag slashed, high speed enabled
Where it is seen	Propeller tips, pumps, valves, rudders	Torpedo & projectile noses
Engineering intent	Avoided at all costs	Deliberately exploited