Hi there, I'm trying to figure out a way to collide flip particles with a shaped container, without having to calculate the full exterior edge of the pool collision.
This is a much smaller version of a very high res sim. I'm trying to work out if the entire inner edge of the pool container shape is necessary to sim, or is it possible to sim just the narrow band surface and still collide with the custom shaped container.
Two examples shown here.. First is a narrow band custom shaped container - Water does not collide.
Second is a narrow band container with custom pool collision shape - Water does collide, but seems unnecessary to simulate the entire surrounding edge.
Is there another way to get the result I'm after, without having to calculate the full exterior of the pool shape?
HIP file attached.
FLIP narrow band collisions with shaped container
733 2 1-
- GlennimusPrime
- Member
- 87 posts
- Joined: 6月 2019
- オフライン
-
- VortexVFX
- Member
- 58 posts
- Joined: 8月 2014
- オフライン
Assuming nothing has changed...
Short answer - unfortunately not.
Long answer - probably, but it requires intimate knowledge of the inner workings of the FLIP solver, and a bunch of bespoke customization of it. The narrowband-enabled solver tracks/advects an interior surface field and unions it each substep with a surface field generated from the particles - you'd need to come up with a way of accurately tracking the volume up to the collision surface, and precisely how to union that with the surface derived from the particles, without causing artefacts or volume loss/gain along the now-open-edged interface between the two.
I've been tinkering with it on and off for years... it's one of a few problems I'd like to come up with a solution to, but I'm not quite there yet
Short answer - unfortunately not.
Long answer - probably, but it requires intimate knowledge of the inner workings of the FLIP solver, and a bunch of bespoke customization of it. The narrowband-enabled solver tracks/advects an interior surface field and unions it each substep with a surface field generated from the particles - you'd need to come up with a way of accurately tracking the volume up to the collision surface, and precisely how to union that with the surface derived from the particles, without causing artefacts or volume loss/gain along the now-open-edged interface between the two.
I've been tinkering with it on and off for years... it's one of a few problems I'd like to come up with a solution to, but I'm not quite there yet

Dan Wood
Vortex VFX Ltd
Vortex VFX Ltd
-
- GlennimusPrime
- Member
- 87 posts
- Joined: 6月 2019
- オフライン
-
- Quick Links