The time has come to churn out something for real rather than endless tutorials and suprise surprise it's some fluid! Except my fluid is vanising… I've fallen at the first hedge as they say.

Collision geometry is really high res but the points just vanish. Point count goes down by half over a couple of hundred frames. If anyone can give me a pointer beyond “make your collision geometry thicker” i would be stoked…. I've attached the hip file and the .obj the the scene reads. Just knowing what's killing the points are going would be a start other wise it's back to the other fluid simulation solution which is a bit counter productive!

cheers,

A>

I'm not a flip expert so I might be wrong, but I think the scene scale is to be blamed for the “leaks”. I'm putting this word into quotes because I believe there are no leaks in the geo, it's just the impression we get from tiny (when compared to over 62 meters tall and 158(!) meters long collision geometry) fluid particles slowly bunching up together before they reach enough volume to pour over the first wall.

I pumped up the particle radius scale from 1.2 to 4 and brought timescale back to 1 (this is optional) and the leak effect disappeared.

However, decreasing the scene scale to something more reasonable is probably a better solution.

Hi Artur,

well my massive chute certainly won't be helping, cheers for that. Do you have a clear idea of what the scale issue was doing? I've always done fluids in realflow as sph sims so flip fluids are new to me really. Could it be something to do with the voxel size of the fluid versus the thickness of collision obstacle? Can I do some sort of visualise on the voxel grid to check that?

Andi, I think I was partially wrong in my post. I just realized that because time scale is set to 10 in your scene, you should also increase number of substeps in flipsolver. This makes collisions resolve properly and gets rid of the “leak” effect. No need to tweak particle scale after all.