hi, i’m experimenting a little with the fluid simulation,
i’m very happi with my resuts but i cant solve a little problem.
let say i want to simulate a not-so-original drop falling in a glass full of water.
i can get nice results if i fill my domain (a cube of course) of water then i add a sphere (the drop) and let the simulation begin
the fluid that fills the domain is calm and start to have reactions just when the drop hit it’s surface.
but using a glass or a cup (as an an obstacle) i make a copy of the glass-mesh to make the fluid inside the glass, and when i begin the simulation i have a sort of small movement:
the fluid seem to adapt itself to thye obstacle and have reactions BEFORE the drop hit its surface…something like wind on the surface of thew sea-
i’m pretty sure this i due to a kind of adaptation of the fluid to the obstacle, but i’m looking for a way to avoid it.
some info:
the fluid baking have been don at 240 resolution, and it SHULD be enought
the cup mesh (and the fluid inside the mesh) have been modelled startin from a 32 side circle…
this is basically unavoidable. The water need to ‘settle’ even if your fluid mesh is an exact copy of the glass. Of course you can play around with various settings to reduce this effect.
What you can try is let the simulation run for a long while until the water settles, then add the drop of water.
This will significantly increase baking time but I don’t think there is another solution.
i bet that having a negative beginning frame will not help, because simulation is calculated a frame after the other, you can not for example bake only frame number 137 , you have to bake from 1 to 137 even if you just want a still render…
what i don’t understan is why a fluid don’t need to “settle” to the domain but need to settle with an obstacle…does this depend from mesh number of vertex?
does a complex (accurate) mesh need less settle?
free-ality: there is such a function. Set the start time to a non-zero positive number, then the first frame of the animation will be ‘pre-rolled’. Note, however, that this means the first frame will take almost forever to calculate.
why a fluid don’t need to “settle” to the domain but need to settle with an obstacle…
Bullx, i think this is because domain is treated differently than an obstacle.
If I understand your problem, you’re having trouble with the fluid-obstacle interaction, correct? Be sure your normals are pointing in the right direction.
i also noticed this problem - the water allready reacts a bit though it didn’t splash. Couldn’t find a way to avoid this. It would be great if they could fix that…something like
if (fluid1 != hit (fluid2_which_is_allready_filled_with_water)) {
don’t do it;
}
Small movements seems to be always present. Some workarounds:
Use a normal mes drop (i.e. not fluid) moving it with Ipo (manually or recorded through game engine) and use Ipo to control Gravity. Set gravity to zero until the drop reaches the fluid surface then graviti to -9.8
BETTER:
Duplicate the Fluid object (that one in the glass)
Set the fluid to Fluid Init Volume
Set the copy of the Fluid to Obstacle InitShell
Use the Ipo Active to deactivate the Obstacle (not the glass but the copy of the fluid in it) when the drop is near to the surface.
One other thing you might try is to increase the real world size of the domain. Larger volumes of water tend to appear more settled, especially at a relatively low resolution. It might not fix everything, but it should help.
It’s easy to start with a motionless fluid…just fill the volume. I have done this many times. But I don’t think this is the only problem the poster is having. The original post says there’s affects before the drop hits the obstacle.