Cavities/Flickering with FLIP solver

[Jassir]

Hi there,

I need to animate the cleaning process of a machine. I use the FLIP solver for this. I like how fast it simulates and how it is able to diplay fine details. But I have some strange behavior near the bottom of the machine, like some kind of flickering. (video:konus_wasser_01.zip) I think it is caused by cavities in the mesh that show up for single frames.
Any idea what I am doing wrong here?

[Ivaylo Katev]

looks like some kind of instability, can you send the scene?

[Jassir]

Sure. That's it: konus_test.zip

[Jassir]

I tried to simulate with more cells (~60 million) but to no avail.
Do you think it would help if I excluded the screw from the simulation?

[Ivaylo Katev]

can't say yet, still busy with other tasks. but you can make some experiments, including to remove the screw or to increase the spf, or decrease the resolution.

[Jassir]

I tried changing the resolution and excluded the screw. But it did not help. I didn't change the spf because I read somewhere I should keep it at 1 if I use the Flip solver. Is that false?

[Ivaylo Katev]

no, flip also needs higher spf for some simulations, it is just not so sensitive as the grid solver

[Jassir]

Ok thanks for that clarification, it is working. A SPF value of 3 seems to be enough to solve this issue.

[Ivaylo Katev]

good to know!
perhaps we must return the default spf to be 2 not 1.

[victorwol]

So what version are you using? becuase I can't do anything with it. It just crashes.

Thanks

[Ivaylo Katev]

the nightlies are broken since april 6, hopefully will be fixed in few days

[Jassir]

I am using the version ...688 (from april 6th).
I had no major issues so far. Only thing I couldn't manage was rendering splashes into the vrayatmosphere render element. It just returned black every time.

[Jassir]

I realized that by increasing the SPF I did not only fix the flickering that I had before. I also changed the speed at which the fluid is emitted. Is that supposed to happen? Is it correct that the only way to adjust the emitting speed is to increase the discharge value in the source object or to decrease the size of the emitters?

[Ivaylo Katev]

as you suppose, only the discharge should control the emitting speed. this is very strange, because the sources produce fixed particle count per second. i made fast test with spf 1 and 2, and the particle count remains exactly the same. can you tell the particle count in both the cases?

[Jassir]

I think the particle count stayed the same, I need to check this. But the emitting speed seemed to change. I have a small source with a discharge of 350. When I simulated with SPF of 1 the particles traveled in a rather straight line for around 4 frames before the visibly dropped due to the gravity. With the same source and discharge and a SPF of 3 they drop a lot 1 or 2. Since I need them to reach the shell of my collision object I have to increase the discharge, right? (The source size needs to stay the same)