zephyr707 Posted July 30, 2012 Share Posted July 30, 2012 hi all, I'm trying to get a liquid stream to flow down a staircase and having some issues using FLIP that hopefully someone can help me with. I'd like the liquid to be sort of like molten metal in its consistency/viscosity, flowing down the steps, but hugging the contours without a lot of "gloop", more like a cohesive steady single stream. I've read through a lot of the help files related to the flip object/solver/sourcing, etc, but am still unable to achieve the desired effect and there are fundamental things that I can't fully grasp yet. I've tried to model everything to scale and build good collision geometry for the stairs, but everything seems very small unit-wise and maybe this is causing issues as well. Particularly with the new fluid source SOP I'm having issues creating particles in the very small emitter. I tried using the pre-built particle fluid emitter. The relationship between particle separation, particle radius scale, and grid scale on the FLIP Object DOP is also a bit of a mystery to me, especially in relation to viscosity. They all seem dependant on one another and I'm not understanding the relationship even after reading the helpful and guides, which recommend certain settings like sqrt(3) for the radius scale and 1.4 for the grid scale, but then it also talks about how these values scale against each other, so I'm not sure which values to use. I tend to just drag sliders around to get the result that I want, but I would like to understand the process better. It makes a bit of sense in the help file, but it doesn't seem to translate for me when simulating and it is frustrating that certain parameters mentioned in the help file do not exist on the node... When I have viscosity set with very small values it all still seems to slide very slowly like molasses on the collision geometry. If I turn off viscosity the particles immediately slide across the surface as if there was no friction. I've played with a lot of settings and substep values and it seems to affect things in different ways, but I can't seem to figure out how to hit the sweet spot for viscosity in relation to fluid motion. It seems that even with extremely low viscosity values it lays a coat of super slow moving particles down and then particles that flow on top of this coat seem to exhibit more logical behavior. Checking the performance monitor, most of the sim time seems to be in the gasviscosity, so something is going on there. Is there a way to check some sort of "absolute" viscosity amount in the details view or is viscosity tightly tied to all the other values? Maybe I need better forces to push this fluid down the stairs, too, right now it is just a uniform force and gravity and the surface is very flat. I've also added a gas surface tension DOP to keep everything tighter, but that doesn't seem to be affecting things too much. I started with FLIP b/c I need to eventually mix two separate streams, but perhaps SPH is a better choice for this slow moving viscous fluid. I've also thought about just using metaballs on a path or with creep for a quick SOP solution. I've attached my lastest scene which is using a source volume emitter in a pipe/tube. A quick flipbook of the effect is attached (there's some strange acceleration on the last 2 steps), as well as an older test that is really slow moving, but hugs the staircase better. Any help/advice would be much appreciated, sorry for the long post! thanks flip_molasses_stairs.mp4.gz for_odforce_stairs_v003.hipnc latest.mp4.gz Quote Link to comment Share on other sites More sharing options...
zephyr707 Posted July 30, 2012 Author Share Posted July 30, 2012 updated file. changed particle scale radius and a few other parameters. This one sims faster, but has a pulsing burst of velocity that causes it to gloop a bit as it goes over a step. for_odforce_stairs_v004.hipnc pulseGloop.mp4.gz Quote Link to comment Share on other sites More sharing options...
zephyr707 Posted August 1, 2012 Author Share Posted August 1, 2012 well, took a break and came back to it, some promising results. took a sidestep into SPH land and back into FLIP. A lot of stuff seems to hinge on that particle radius scale, have a better grasp on it now, but still don't fully grok it. oh well, the fluid motion is good enough for a prototype, will have to figure out better controls for keeping it from pooling up too much after the top 2 steps. Will maybe use blocking geometry or pull it towards a path. reading over this help card on simulating liquids multiple times helped a lot. There's a lot of good tips in the help files, but a lot of it seems squirreled in various places. cascades.mp4.gz 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.