Turbulence in liquid at rest

[GrahamMac]

I have a scene with a short jet of liquid which 1/3 fills a martini glass and then over 200 frames is allowed to come to rest. Whilst coming to rest there is a gradual turbulent boiling motion in the liquid that grows in amplitude over time.
I would love to discuss any ideas on what can cause this as I've had no luck yet with my tests.
Settings are shown in the image attached.

[Ivaylo Katev]

making settled liquid with flat surface is a hard task not only with phoenix
but, it's not impossible. in one of the examples (paints.max) the liquid becomes flat after the container is filled from the pipes. unfortunatelly there is no guaranee that the same settings will work in other scenes. so, what is the general line:
- use fixed spf mode (equal upper and lower range) instead the adaptive one, that is by default
- calculate your desired spf ratio starting from the paints.max scene, with the following in mind: if your cell size is twice smaller that the cell size of the paints.max, you have to double the spf. if you gravity is twice stronger, you have to double the spf. with other words, you have to observe the ratio spf*cell_size/gravity to be like in the paints.max sample.
- the conservation plays really important role for the liquid's stability. first rule is that it should be at least 20, but making it bigger not always produces better result. my recomendation is to export the velocity channel, to set the preview to show the liquid between 0.1-0.5 (the surface) and the velocity with AND condition. in the most cases the velocity of the surface is directed in pozitive z direction (rest velocity). changing the conservation, the rest velocity changes its magnitude and direction, tending to alternate in 20 to 100 conservation units. the best chances for stable sorface you have when the rest velocity is as small as possible.

[GrahamMac]

Your suggestions have cleared up the turbulence at rest very nicely!
I'm using SPF upper/lower = 11 / 11
I switched on velocity and set it to 1 in the source and turned on velocity in the simulators output rollout.
With velocity preview on, including the AND condition, nothing meaningful is displayed in the viewport until I change the velocity preview high and low to around 4000 / 20. Does this sound about right?

I'm not sure what the velocity is doing to fix the problem though. Its not covered much in the help files but it does say that velocity is used in the simulation even if unchecked.

[Ivaylo Katev]

yes, the velocity is an essential part of the simulation
i forgot to say that you need to change the velocity preview range, but you found it alone as i see.

[GrahamMac]

Thanks again for your help. I think it is very close to working but I have hit another snag:
I need to have more liquid pouring into the martini glass and have found that the turbulence problems are back. (see image)
Increasing conservation quality or SPF upper/lower values (keeping equal) help a little but does not seem enough. Rather than pushing these values through the roof is there a better approach?

The turbulence is strong in a small area closest to where the last remnants of fluid are sliding down the inner surface of the glass into the liquid.
These remnants are below the 0.5 surface level so do not render however even after they are gone the turbulence does not settle down over time.

Additionally - appreciating how velocity is essential to the simulation, what do the velocity controls intrinsically do? (since even when off, the velocity must be used anyway)

[Ivaylo Katev]

try to decrease the gravity, the martini glass requires very high spf value if real scale is used, so decreasing the gravity can help.

[GrahamMac]

Thank you Ivaylo, I've been busy experimenting with the parameters you suggested and am now getting some good results, although I still find the turbulence is not completely eliminated. I noticed a curiosity with it though: Looking closely at the turbulent area on the surface I can often see a line of turbulence within the volume of the liquid reaching down and touching a fixed point on the inner surface of the glass/container. It looks like the turbulence is anchored there for some reason whilst the surface turbulence is free to wander around a little.
Something else you might be interested in as a workflow idea, which I haven't seen mentioned anywhere yet: To funnel the discharged fluid through a tube which adds to the control one gets with the look of the pouring stream. Once you find the right discharge rate so you know a container wont overfill or under-fill the apparent flow rate and constriction can be adjusted with a taper modifier on the tube.