Bernoulli principle: Venturi tube

Water speeds up in the throat and its pressure falls; read it on five glass standpipes.

Switch to 2D

Presets

Playback

Venturi bench

Camera

Layers

Quality

Model check

The science

What is modelled

Water flows by gravity from a header tank through a glass Venturi tube (32 mm pipe, a cone down to the throat, a long gentle diffuser back to 32 mm) and out through a valve into a tub. Five glass standpipes (P1 to P5) show the static pressure along the tube as a height of water. The tank level, the throat diameter, the valve and the pipe friction are adjustable.

Equations

  • Continuity for steady incompressible flow: Q = v A is the same at every section.
  • Bernoulli with losses along the flow: h + v²/2g + losses = H, where h = p/(ρg) is the static head (gauge) and H the water level above the pipe axis. At the throat the speed is high, so the static head is low: h1 - h2 = (v2² - v1²)/2g without losses.
  • Losses: entrance 0.5 v1²/2g; friction ∫ f/D · v²/2g dx (Darcy); the diffuser keeps only 85 % of the kinetic head it recovers (0.15 (v2² - v²)/2g is lost); outlet valve Kv v²/2g; the exit jet carries v²/2g away.
  • The flow follows from H = R Q², the sum of all these terms.
  • Venturi meter: Q = Cd A2 √(2g (h1 - h2) / (1 - (A2/A1)²)).
  • Cavitation: the pressure cannot fall below the vapour pressure of water (2.3 kPa absolute, which is -10.1 m of gauge head at 20 C). If the throat would go lower, the flow is limited so that the throat sits at that limit.

Checks run in the page

  • Integrating the momentum equation along the tube step by step gives the same static heads at all five taps as the closed form.
  • The Venturi-meter formula applied to the readings of P1 and P3 returns the true flow with a discharge coefficient just under 1.
  • The losses plus the exit kinetic head add up to the tank head (energy balance).
  • With no losses the exit speed equals Torricelli's sqrt(2gH) and the inlet-to-throat head drop equals the Bernoulli value.

Honest simplifications

  • Steady, plug flow (one velocity per section); the real velocity profile, the boundary layer and turbulence are not modelled. The friction factor is a constant you choose, not computed from Reynolds number and roughness.
  • The loss coefficients (0.5 entrance, 0.15 diffuser) are typical textbook values, not measurements of this apparatus. The sudden-contraction loss at the cone is neglected.
  • A real open standpipe cannot show a negative gauge pressure: it would suck air in. The model still reports the head; the standpipe is drawn empty. Opening the valve fully produces this suction at the throat.
  • Cavitation is only a limit on the flow. Real cavitating flow is two-phase and loses pressure recovery in the diffuser; this model does not simulate it. Dissolved gas makes real cavitation start earlier.
  • The tank level is held constant (an overflow keeps it full); the tracers are drawn slowed down.

What to try

  • Close the valve (resistance 300): the heads are high and nearly equal, because there is little flow.
  • Open the valve (resistance 10): the throat static head becomes negative, and a smaller throat makes it worse.
  • Use the cavitation preset: the throat reaches -10.1 m and the flow stops increasing.
Loading model...
Loading Blender model...
Head along the tube
Flow vs throat diameter
Throat pressure vs throat diameter
Made with 🟠 Blender · Three.js