Treat interface resolution as a coupled space-and-time problem. Refining the mesh without reducing timestep—or reducing timestep around a badly under-resolved interface—often produces little benefit.
Problem: the phase interface loses shape or becomes unstable
Typical VOF problems include excessive interface thickness, unphysical droplets, jagged free surfaces, volume-fraction overshoot/undershoot, waves that decay too quickly, waves that amplify, violent startup transients, or instability near outlets and moving boundaries.
The right fix depends on whether the problem is primarily transport resolution, mesh topology, boundary reflection, initialization or a genuinely violent physical breakup process.
Classify the symptom first
Interface smearing
Volume fraction transitions across too many cells. Suspect insufficient spatial resolution, excessive numerical diffusion or overly large timestep.
Interface chatter/noise
Small-scale oscillations appear along an otherwise smooth surface. Check mesh anisotropy, timestep and interface-compression aggressiveness.
Wave decay
Amplitude decreases too quickly with distance/time. Numerical diffusion, coarse mesh or unsuitable wave-boundary treatment may dominate.
Wave growth/reflection
Amplitude increases or standing waves form unexpectedly. Check outlet reflection, damping zones, domain length and boundary forcing.
Startup explosion
The first few timesteps generate violent velocities or pressure spikes. Initialization may be inconsistent with hydrostatic pressure, gravity or the imposed boundary motion.
Fragmentation
Droplets/bubbles appear where physics does not support breakup. Determine whether they are numerical artifacts or under-resolved real structures.
Likely cause groups
Timestep
Interface travels too far per step relative to the local cell size, causing poor advection and instability.
Mesh resolution
The wavelength, film thickness, jet, gap or curvature radius is represented by too few cells.
Mesh alignment/quality
Highly skewed cells or abrupt size changes near the interface increase advection error.
Initialization
Volume fraction and pressure fields do not satisfy gravity or the starting free-surface shape.
Boundary conditions
Reflections, incompatible inlet phase fractions, poor outlet placement or insufficient damping contaminate the domain.
Interface numerics
Compression and reconstruction settings can trade sharpness against robustness; aggressive settings are not a substitute for resolution.
How to diagnose a bad VOF interface
- Plot local Courant number around the interface.
Use the actual interface-region velocity and cell size, not only a domain-average estimate.
- Count cells across the feature you need to preserve.
Check wavelength, film thickness, jet diameter, gap height or expected interface curvature against local mesh size.
- Visualize volume fraction with multiple contours.
Inspect 0.1, 0.5 and 0.9 iso-levels to see whether the interface is thickening or fragmenting.
- Check mass/phase conservation.
Unexpected loss or gain of a phase can indicate transport/boundary problems even when the visual result looks plausible.
- Compare upstream and downstream wave signals.
Amplitude and phase history can reveal reflection or numerical damping better than a single snapshot.
- Reduce timestep and refine only the interface zone.
A controlled sensitivity test shows whether the solution is limited by space, time or both.
Estimate the timestep scale
Use local velocity and cell size to screen the Courant number before running another VOF case.
Fix order
- Correct phase initialization and hydrostatic consistency.
Start from a physically coherent pressure and volume-fraction field.
- Reduce timestep to resolve interface motion.
Use the local restrictive interface region rather than a comfortable domain-average CFL.
- Refine where the interface travels.
Use local or adaptive refinement when the free surface moves through a large region.
- Repair reflections and boundary placement.
Increase domain length, improve outlet treatment or use damping where wave energy should leave the domain.
- Then tune interface-capturing numerics.
Adjust compression/reconstruction only after space, time and boundaries are credible.
Common mistakes
- Refining the interface mesh while leaving the timestep unchanged.
- Using a global Courant number that hides a highly restrictive local cell.
- Starting with a flat pressure field in a gravity-dominated hydrostatic problem.
- Placing an outlet too close to the wave/free-surface interaction region.
- Interpreting every small droplet as physical breakup without a mesh/time sensitivity test.
- Using aggressive interface compression to compensate for an under-resolved mesh.
Related Abecator resources
Need help stabilizing a VOF/free-surface model?
Share a non-confidential description of the geometry, phases, mesh scale, timestep and observed interface behavior.