tebling Posted September 20, 2020 Share Posted September 20, 2020 Hi folks, first post here. I've been doing some tests with moving FLIP containers, and for the life of me I can't seem to fix this issue with particles disappearing into the collision mesh. This example uses 10 substeps on the DOP network just to rule that out as an issue (though it doesn't seem to matter whether substeps are there or on the solver itself), and the collision mesh is run through a collision source node, which does generate reasonable looking collision velocities (viewed using the FLIP source guide). I've tried reseeding and turning on particle separation with similar results. Scene file attached - would really appreciate some tips, thank you! Water Wheel v_01.hiplc Quote Link to comment Share on other sites More sharing options...
bunker Posted September 21, 2020 Share Posted September 21, 2020 https://www.sidefx.com/forum/topic/49968/?page=1#post-225812 Quote Link to comment Share on other sites More sharing options...
Atom Posted September 21, 2020 Share Posted September 21, 2020 Here's my tweak of your scene. I basically reset everything to the defaults. ap_water_wheel_v_01.hiplc 2 Quote Link to comment Share on other sites More sharing options...
tebling Posted September 21, 2020 Author Share Posted September 21, 2020 Thanks @bunker for the link, that was a great read and I learned a ton from it, but ultimately none of the tricks in there solved my problem. Many thanks @Atom for the work you put into getting my scene to work! I meticulously went through my own scene and set my values to yours, testing along the way in order to debug. Of course when I found the issue I smacked my head - the problem was that I'd set the time scale on the DOP node to 0.5, and I guess what happens is that doing so makes the Collision Source generated velocities incorrect. You guys are awesome - thanks again! Quote Link to comment Share on other sites More sharing options...
bunker Posted September 21, 2020 Share Posted September 21, 2020 Also don't expect FLIP to preserve volume, it just doesn't do that. The particles are added or removed based on the surface field - that's only a rough estimate and the volume loss becomes worse the more frames and/or substeps you have. This old masterclass explains some of the concepts of FLIP https://vimeo.com/15571778 3 Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.