Hi,

I have some FLIP questions, any help is appreciated

1- In particles reseeding, how is "Particle reseeding -> Particles per Voxel“ useful? meaning, what difference does it make to set it to 4,8 or 16?

2- ”FLIP Object->Grid Scale“ and ”Particle reseeding->Particles Per Voxel" seem to define the same thing? i.e a bigger grid will have more particles per voxel, so why redefine the number of particles per voxel again?

3- What does mean that particles are “unresolved” ?

4- I understand that the “pressure projection” solve is the process that changes particles velocities to compensate for pressure differences (i.e. move particles from high to low pressure). My question: if the number of particles per voxel can change freely (i.e. particles reseeding, voxel size, …) what really defines pressure? i.e. what makes some areas of fluid have more pressure than others?

5- According to docs, Particles reseeding will increase spawn particles if they are below a certain count per voxel and vise versa. My question: what makes particles diseapear in the first place? I understand they are moving, but so what? If they leave a voxel then the voxel becomes empty and it's even “wrong” to create particles just to meet the minimum particles count per voxel? The same question applies on having “too” many particles per voxel, what this happens in the first place and what does it mean?

Thanks

Anyone ? Answering a single question will also be appreciated

1. this should usually be set to gridscale^3 to represent the same point density within voxel as flip object's default setting
but it's really up to you to set it to the density you wish to try to maintain

2. they are related, however they don't have to be kept at numbers that represent the same density, while Flip Objects grid scale defines the actual size of the voxel based on particle separation value
reseeding allows you to oversample or undersample the actual particle density based on it's own settings

3. I guess docs explain this quite well

4. pressure is fully independent of point count or point density per voxel as it's solved purely on voxels, so the main thing that defines the pressure is velocity, surface and collision field, like in smoke sim, pressure projection tries to make velocity field non-divergent/incompressible, so whatever causes divergence creates the pressure that is used to negate that effect

5. this is not just arbitrary seeding/deleting of points in voxels
it depends on your surface field, any particles fully within surface are considered inside of the fluid and therefore should be safe to enforce optimal point density per voxel
anything that's within the boundary voxels may as well need at least minimum number of particles to properly represent that boundary especially during meshing, but may as well need some further oversampling to ensure smoother results
what makes particles over/undersaturate voxels is their velocity, fast collisions as well as overall motion as especially in thin sheets of fluid there may not be enough particles to describe expanding sheet as it's only represented by finite amount of particles that is much smaller than number of water molecules in reality
particle saparation can also help gradually equalize point density per voxel if you wish to keep reseeding off

Thanks Tomas so much for your reply Do you think I need to understand the math/physics of FLIP simulation in order to better control/understand the sim? or experience and testing are enough?

I have other questions, any help is appreciated

A) Are there any advantages for having particles per voxel more than gridscale^3? (i.e. changing the particle reseeding to more than 8)? The logical answer is “smoother” sim, but I am questioning: having more particles than gridscale^3 seems like squeezing too many particles in one voxel that should eventually spread up to maintain their particle separation?

B) Liquid Surface : you discussed that voxels inside the surface should have enough particles to represent the liquid. My question: who defines liquid surface in the first place? From my simple tests, I think it's defined by the liquid emitter and then it changes to accommodate for particles leaving the surface? (i.e. in case of a splash, many particles leave the surface, hence the solver somehow changes the surface to encompass the new particles?)

Thanks,

catchyid
seems like squeezing too many particles in one voxel that should eventually spread up to maintain their particle separation

A subtle point about FLIP sims is that particles are just markers. They don't actually take any space themselves. This is why spawning particles inside a tank doesn't cause the water level to rise.

catchyid
My question: who defines liquid surface in the first place?

It is a chicken and egg question :>

The particles define the surface; then the surface is used to define the particles. In standard FLIP, each frame the surface is created fresh from the particles and isn't used on the successive frame. But during the frame's computation it is used for things like determining if particles are deep in the fluid or not.

Thanks Jeff for your help I “think” I should read more about the physics of FLIP simulation to better understand how things work