I did a mid resolution flip simulation:

https://youtu.be/5ROyYxwVSxQ [youtu.be]

With which I was pretty happy.

So then I let it sim all night at a high resolution:

https://youtu.be/flnZufED2qY [youtu.be]

Take a look at frame 140. In the mid res version there is a nice build up of the water close to the walls at the right. This is completely lost in the high res version.

Is there some parameters that I could adjust to prevent this ?

you probably need higher particle radius scale in that case, increasing the resolution causing your particles to come close to each other ( i.e your particle separation) and you are seeing this nice sharp simulation result , with low res the separation is high so they are keeping more distance.

Thanks, so what value would you suggest ? maybe go from the default 1.2 to 2 ?

One thing though, is that according to the popup help, setting this parameter to a higher value will result in less surface detail.

So would you still try this ? What about the similar-sounding “Collision Separation” or “Apply Particle Separation” in the flip solver ?

I'd be more than happy to test this myself but as you can imagine, it'll take weeks to make comparisons with just a few combinations.

sure, you wont be able to get detailed surface results with higher particle radius (that is why its not a preferred way if you dont want to compromise your details). What version of houdini are you using and what is your system configuration, this looks fairly straight forward sim and shouldn't take over night to calculate as long as you are on a very slow system.

I would increase the particle count in the source if I want the fluid to collect at the wall and add a bit friction to the collision geometries , make sure you are using vdb's for collision, they ll make your sims faster.

I'm using 16.0.541

i7-6800k 3.4 ghz 32 gb ram

On the slower frames it takes about 2 mins 30 secs per frame.


The particle count in the source is directly dependent to the particle separation right ? So when I did the high res sim the particle count in the source increased accordingly…

I'm gonna try your suggestions on the collision, thanks a lot.

u could do upres for your low sim by advecting high res sim to low res, or try to find a better view by changing parameters like Separation, decreasing Velocity Smoothing and Surface Extrapolation, and switch Collision Detection to Particle, all of these parameters in FLIP Solver, maybe it will help