Dripping Faucet Chaos Lab: Tate's Law, Period Doubling and Bifurcation | mysimulator.uk

A Blender-modelled tap above a sink: surface tension, inflow and a recoiling remnant turn a regular drip into period doubling and chaos as the flow rises.

Switch to 2D

Presets

Playback

Parameters

Camera

Layers

Quality

Model check

The science

What is modelled

A tap dripping into a sink. The water hanging from the nozzle is treated as one lump of mass m(t) that grows as water flows in and is held up by surface tension, which acts like a spring. When the lump stretches to a critical length a drop breaks away and a small remnant recoils upward. Whether the drips come regularly, in two-beat or three-beat patterns, or chaotically depends on how fast the water flows, as in the classic leaky-faucet experiments of the 1980s.

Equations

  • Inflow and momentum, new water entering at rest: dm/dt = ρQ, m v′ = m g − k x − b v − (dm/dt) v, with x the stretch of the lump below its rest position and v = x′.
  • Tate's law fixes the spring: the static drop weight is mc g = 2π R γ f with R the nozzle radius, γ = 72 mN/m for water and f ≈ 0.6 the Harkins-Brown correction for the part of the drop that stays on the nozzle. Break-off happens at xc = R, so k = 2π γ f (about 0.27 N/m).
  • Damping: b = 2 ζ √(k mc), where ζ is the damping ratio of the hanging drop (viscosity).
  • Break-off rule: the drop carries away m − mr, the remnant is mr = mmin(1 + 5 max(0, v)/v0) with mmin = 0.35 mc (a faster-moving neck pinches off higher up and leaves more water behind), keeps its velocity and restarts at 0.85 xc. This memory of the previous drop is what makes the next interval depend on the last.
  • Integration: fourth-order Runge-Kutta with 0.2 ms steps; break-off is detected at the end of a step, so each interval is accurate to about one step (0.06 % of a 350 ms interval). The falling drops are free fall, y = ½ g t², with no air drag.

Checks run in the page

  • In the slow, damped limit the break-off mass equals the Tate drop weight to within 1.5 %, and the resulting volume is close to a real medical-dropper drop.
  • Water is conserved: initial mass plus inflow minus the remnant minus all drops equals zero to 1e-12 kg over 40 s of chaotic dripping.
  • The hanging drop's small-amplitude ringing period matches the damped-oscillator formula.
  • Reducing the step four times changes the steady-drip intervals by no more than one coarse step (0.25 ms).
  • Two runs started 1e-9 m apart diverge in the chaotic regime and stay together in the regular one.

Honest simplifications

  • This is a lumped mass-spring model of the Shaw and Martien kind, not a fluid-dynamics solution of the Young-Laplace and Navier-Stokes equations. The parameters 0.35, 5 and 0.85 in the break-off rule are chosen to give the qualitative route (regular, period-2, period-3, chaos) seen in experiments; they are not measured for water.
  • Real chaos of a faucet appears at flow rates and nozzle sizes that depend on the apparatus; the numbers on the flow slider belong to this model. Only the shape of the story (a bifurcation sequence with flow rate) is claimed.
  • The drawn hanging drop is a schematic ellipsoid whose volume follows the model mass, not the solution of the pendant-drop equation; satellite droplets and drop oscillations are not modelled, and very small intervals in the chaotic regime stand for the small fragments that real breakups leave.
  • Fluid properties are those of water at 20 C; the nozzle is perfectly wetted at its outer edge.

What to try

  • Press Bifurcation diagram and read off where the single line splits into branches; then set the flow slider to a branching value and compare the interval graph.
  • Watch the return map (second graph): a regular drip is a single point, a two-beat drip is two points, a chaotic drip draws a curve, which is the signature of deterministic chaos rather than noise.
  • Make the nozzle thicker or the damping higher and see where the chaotic window moves.
Loading model...
Loading Blender model...
Time between drops (ms)
Return map: interval n+1 against interval n
Bifurcation diagram: intervals against flow rate
Made with 🟠 Blender · Three.js