wrong water behavior (help)

[vhalldez]

I had been dificults to find a right behavior for water. The water was spreading out  a lot...but I want that it fall more together. (as follow image)

Some one know how to fix this?

 

[catchyid]

hello,

you can try some ideas (or mix them):

-add some viscosity to make it stick more together, a small value bet 0.1 or 1.0 *might" fix it

-if your bottle is not transparent, then you don't need to sim the water going down the bottle which I guess creates this erratic velocity, so if this is the case then just add an emitter at the top of the bottle with an initial vel toward the cup, this might give you a smoother velocity

-try to a different kernel, i think this is the splashy kernel, try the swirly one 

-add some damping to the particles (pop damp)

-after the sim , i.e. post sim, kill isolated particles (hmm, won't fix this problem much, but if u got lots of spray like particles this will fix it)

 

[Stephen Moroz]

Hey, I'm also doing a water pour sim at the moment, and here are a couple things I've found:

- Firstly, you can't sim it to real-world scale because the values will all get too small when solving and this can cause errors. For mine, I'm using 5x scale and it seems to be working without issues.

- When simulating fluid at a bigger scale, you need to turn up the viscosity to compensate so that it still looks the correct scale. (I think most of the flip solver is scale independent but don't quote me on that!)

- To fix the particles spreading out too much, you need to turn on surface tension. Even a very low value should stop it spreading out so much, but you can play around with it and see what looks best.

- Finally, to reiterate what Khaled said, I recommend swirly kernel on the flip solver, since it is better for simulating smaller things. You can read more about splashy vs swirly kernel here: http://www.sidefx.com/docs/houdini/nodes/dop/flipsolver.html#splashy-vs-swirly-kernel

Hope that's helpful!  

[vhalldez]

thank u a lot guys...

To change karnel for swirly and add surface tension...solver the problem.

Add vicosity i had tried before, and dont work hor I wish,,,,now it work too..thak u again.

[nigelgardiner]

Hi Valdez

I've seen it mentioned a number of times on the forum that building sims at oversized scales is the way to go but I'm not so sure this is correct. At least for things like pouring a bottle of liquid into a glass.

Having played with Flip fluids for a while now I think it is possible to run at the correct scale and it gives results that are more realistic than large scale sims. It does involve a lot of testing but no where near as much as sph fluids.

It seems to me that unless you are simulating a viscous fluid you don't need to invoke the viscosity solve.  Swirly kernel is the best option. Divergence is going to be necessary to stop shrinkage if your fluid is settling. Surface tension is going to be a useful force as when looking at small scale fluid it has a big impact. Enforcing air incompressibility on the solver is also a good option. The speed at which your pour develops and the shape of the pourer are important.

When looking at a sim I think the surface guide is a better indication of what's happening at small scales than the particles as it really is all about the surface of the fluid.

I think you can get good results with as little as 500000 particles if you are pouring say 500ml of liquid from a bottle to a glass.

 

Hope this is useful.

cheers

Nigel.

 

[nigelgardiner]

Here's a fail from the other day, the pour is too fast, but this is at actual bottle and glass scale.

v11e .mp4

[nigelgardiner]

Oh and should have said sub steps are up around 6 or 7 on the flip solver and total sim time is around 12 hours on a dual hex core @ 3.3ghz (or 3.5 i forget) . Memory usage is only around 15-20 gig I think.