Emitting from polygons only in normal direction?

[ben_hamburg]

Hi there,

Finally got another project with fluids again...
My problem this time: I have an object and selected some faces on it, from where I want it to emit liquid (sort of like "sweating" it).
So I set the selected polygons to ID 2, created the PHX sim and a liquid source and set it to "brush" and chose polygon ID 2.
It does emit from selected polygons now, but it doesn´t only emit them in the right direction (facing away from the object), aka along its normals, but also in the opposite direction.
How can I change that?

[Ivaylo Katev]

can you post an image?

[ben_hamburg]

I already went another way, so I don´t have a screenshot at hands....I just wanted to know if that is normal behaviour or just something I´m doing wrong in my setup:
For exmple: The emitting geometry is also a collision object for the liquid, so it might just be pushed through the collision object or something like that...

[ivaylo.ivanov]

The brush mode should not actually discharge fluid with a force. It should directly set the fluid content. Maybe you have force coming from somewhere else that is pushing the fluid in opposite direction.

[Ivaylo Katev]

i suppose the brush option is ignored because the object is solid.

[ben_hamburg]

The Emitter object is actually just some faces from a displaced plane, so it´s not solid.
I think I pinned down the problem though: The Collision object was also just a plane with displacement, I just added a shell modifier and now it seems to work.
I guess gravity was just pulling the liquid through the collision object and because the emitter was so close to the collision object, it looked like it was emitting from the wrong side...
So I guess collision objects should be solid? I also tried unchecking "solid" in the phoenix properties of the collision object, but that didn´t help...
I don´t mind putting a shell modifier on top, I´m just asking for future reference...

[Ivaylo Katev]

i think i found what causes the confusion. the polygon id option works only for solid objects, and the plane is solid by default, but is not caught by the simulation due to the inscribed voxelization mode. however, the source is still trying to use it and this causes the strange behavior. you have solved the problem using shell modifier, that is the same as if you just switch to circumscribed voxelization.

[ben_hamburg]

Ah, thanks for clearing that up!

I got another issue now though, maybe you can help here aswell:

I want the liquid to slowly build up (slow, because the liquid has a high viscosity, it should still build up fast enough, I don´t want the shot to be too long) and then flow down the surface, so I animated the discharge over 25 frames from 0 to 50. I did some low grid sims to test the dynamics and it looked fine, building up and then flowing down as desired.
But as I inreased the grid resolution, I got lots of ripples at the places where the emitter objecs are.
Just as a test I deleted the discharge animation, and now I got nor ripples, but the liquid is also not building up as well as before (I haven´t let the sim run the full range, but it seems like it so far).

So my question would be:

How do I get that build up without the ripples?

I already inreased the step size to 8, but still get a lot of ripples if I animate the discharge.


So what would be the correct way to get that slow

[ben_hamburg]

hm, might be the amount of liquid coming out of the emitters in that amount of time is just too much to make a "clean" build up, I think the "ripples" I am seeing are just splashes due to the speed the liquid is beeing emitted...
I´ll have to do some more testing to confirm this and then find a workaround to get the amount of liquid I want in the time I want, without putting too much pressure on it to splash around like crazy...

[Ivaylo Katev]

in the most cases when increasing the resolution produces problems i recommend to increase the spf with the same proportion as the cell size. however you says that already did that (under step size you mean spf right?), and this does not help. perhaps if you attach some image or short movie this can give me an idea.

[ben_hamburg]

I think it is the splashes from the amount I am discharging.

There still is something weird going around frame 200. I animated the discharge from frame 150-200.

Here is also a screenshot of the Settings:



And of the emitter (pink polygons):



EDIT: I just noticed the emitter polygons were flipped, facing down towards the collision object.
I´ll try a quick resim to check if it runs better with correct normals...

[ben_hamburg]

goddamnit...

flipped normals seem to have been the problem.
While I´m still struggling with finding the correct settings for every project, wrong geometry due to beeing impatient still accounts for most of the problems while testing...

I just wish I had more time in between projects to do some proper testing.
I´d love to make one of these comparison charts I´ve seen for other programs like fume FX or realflow, where all settings are beeing played through with little videos showing the different results...

[Ivaylo Katev]

there is a warning when you have flipped geometry, but it does not work if the geometry is relatively wide and open

[ben_hamburg]

I see...that´s good to know, but I´ll just keep in mind that if the sim isn´t behaving as expected, I should always first check my geometry....