Prelimina
All applications
Application / 07 · Tanks & prescribed motion

Check pressure in a sloshing tank.

Does the simulated wall pressure follow a sequence of sloshing impacts in a rolling tank, and do the impacts arrive on time?

Availability: In development

Inputs
A closed tank, 900 × 508 × 62 mm inside, with 93 mm of water, rolled about the centre of its floor by the measured roll history (±4°, period 1.63 s). Water properties and gravity.
Variables to explore
Grid spacing (6, 4 and 3 mm), the time-step limit (CFL 2 and 0.4), and the particle-transfer settings.
Keep the comparison consistent
The tank, fill, roll history and sensor position of SPHERIC benchmark Test 10: pressure sensor S1 on the left wall, at the still-water level.
Outputs
The S1 pressure history and the tank motion, exported as CSV, and the live fluid view.
Comparison with measurements
All four impacts reach S1 within 35 ms of the measurement. Pressure impulse at S1 over each impact, from 0.4 s before to 0.3 s after the measured peak, against the reference run: −2 % for the first impact, then +35 %, +47 % and +60 %. On 6, 4 and 3 mm grids the first impulse stays between −12 % and +2 %; the later three range from +29 % to +72 %, and two identical 4 mm runs differ by up to 25 percentage points.
Physical assumptions
A single liquid; the air is not simulated, so the air trapped at the impacts is absent and the water closes freely. Each pressure is the average over a 6 mm sphere of water at the sensor, not a flush wall transducer. The tank rolls as a rigid body in a still domain, driven by the measured roll rate.
Scope
One sensor and one published experiment. The first impact gives a single-step pressure spike of up to 550 mbar against a measured 37 ± 7 mbar, and the spike grows as the grid is refined, so impulses are compared rather than peaks. From the second impact on, the pressure after each impact sits about 3 mbar above the measurement under every setting tried, so the later impulses come out 30–70 % high; the air that the experiment traps at these impacts, which is not simulated, is the leading candidate. From the second impact on, a few hundred particles hang as spray above the water.
APPLICATION / 07Simulation
Side view of a long, closed tank rolling a few degrees each way. The shallow water runs to one end and slams into the left wall, sending a sheet up the wall to the lid, then runs back; four times in all. Below it, the simulated pressure at the wall sensor S1 draws in step over the measured trace, each impact lining up with a measured one.
A Prelimina FLIP run with the record settings: the closed tank rolls in a still domain, driven by the measured roll history. Side view coloured by speed (0–1.5 m/s); isolated spray particles, about 0.5 % of the water, are not drawn. The pressure at S1 draws in step over the measured reference run; both use a 10 ms moving average. Development build 1b766fd6.
Top: the pressure at S1 over 8.35 s. The Prelimina curve rises with each of the four measured impacts and then sits a few millibar above the measured pressure while the water stays against the wall. Bottom: four close-ups of the raw signals around each impact, with the spread of the peaks over 102 repeated experiments as a band. The first simulated impact shows a single tall spike; the later peaks fall near or above the band.
Pressure at sensor S1 on the left wall, at the still-water level. Top: both signals with a 10 ms moving average. Bottom: raw signals around each impact against the peak mean ± 1 standard deviation of 102 repeated runs. Experiment: SPHERIC benchmark Test 10 (Souto-Iglesias and Botia-Vera), lateral impact, water, 93 mm fill. Prelimina FLIP with density projection, no PBD and wall boundary damping one cell wide, 4 mm grid, CFL 0.4, closed tank. Development build 1b766fd6.
Discuss your case