Problems simulating very small fluid volumes (mm‑scale) – is this realistic with SPlisHSPlasH?

[chirumla]

Hi everyone,
I am trying to simulate very small fluid volumes in the millimeter range squeezed between two plates (single droplets that are only a few mm³). The setup includes:

• particleRadius around 0.00007–0.0001
• DFSPH (method 4)
• high viscosity (oil-like)
• Becker & Teschner 2007 surface tension
• moving rigid plates (PBD slider joints)

The issues I encounter:

• the droplet slowly breaks apart over time
• sometimes it bounces off surfaces even with strong adhesion
• surface tension seems hard to balance
• performance becomes very slow because of the small particle radius
• I am not sure if SPlisHSPlasH is numerically stable at such small scales
• the droplet keeps a distance from the surface and it starts to "dance"

Before I continue tuning parameters, I would like to know:

• Is simulating mm‑scale droplets with strong viscosity and surface tension physically realistic with SPlisHSPlasH?
• Or is it recommended to scale the geometry up and run the simulation at a larger scale?
  I will attach my full scene file below.
• do you have any recommendations regarding setup?

Thanks for any advice!
Christian
klein3.json