Generate smoke/particles from Pflow

[ben_hamburg]

The particles are emitted and then move between two surfaces. What I can see in the velocity preview might just be the air beeing pushed around between those surfaces. After simulating some frames I can see the grid stays red for all those frames, whichs, as far as I understood, means that there is nothing in it...

[Ivaylo Katev]

yes, it may be the geometry influence...

[ben_hamburg]

Argh, I hate this sort of stuff...I now merged only the particle flow and it´s influences into an empty scene and now it works...I´ll merge the whole scene into an empty one and maybe I´ll get lucky and it all works...otherwise I´ll have to take out one element at a time to find out whats causing this.

[ben_hamburg]

Well, I rebuilt the scene from scratch and now it seems to work...

[ben_hamburg]

So I´m testing right now with a pretty low resolution (only 10000 cells) and I´m starting to see grid artefacts (the drag particles are forming square forms at pretty much the grid cell size).
Do those artefacts dissappear with higher resolutions or do I need to do anything else here to get rid of the artefacts?

[Ivaylo Katev]

the artifacts will disappear if the particle density is the same, but better to increase the dynamics in order to scramble the particles.

[ben_hamburg]

I didn´t quite get that...

What do you mean by "the particle density is the same"? The amount of particles emitted and the total cell count?
And by "increase the dynamic" you mean increase the default "dynamics" value of "1" in the dynamics tab?

Also what is a good particle count? I´m already emitting 2 million drag particles and I feel I can only see a fraction of that in the viewport preview.
Or is there a viewport reduction going on?

[Ivaylo Katev]

with dynamics i mean the velocity in the grid, actually i'm not sure what you are trying to do and may be this is not appropriate.
about the reduction - it is controlled manually, if the reduction is 0 there is no reduction.

[ben_hamburg]

Hm, well, I basically have some debris moving around (via pflow) and want some secondary finer debris/dust, which should come from phoenix.
I´m just wondering why 2 mill particles look like 2000.
I haven´t worked with drag particles yet, so I´m not quite sure how to increase the amount.
Do I simply increase the particle count (like I did)? What do "Velocity" and "Discharge" do when using a pflow source for emitting drag particles?
Because like I said: I bareley noticed a difference between 10000 and 5 mill particles in my setup and according to the manual the number behind the "particles" checkbox should indicate the birth ratio...?
And discharge didn´t seem to do much either, I couldn´t see a difference between 10 and 1000.
Then again, I still don´t trust my scene...maybe it´s still broken and al these value would behave completely different in an empty scene like the one I tested...

[Ivaylo Katev]

can you post an image, rendered or from the viewport?
thers is some possibility to be a bug, if for some reason the particles are born in the same position, they will continue to move together, because they get the same velocity, in this case they can look low in count. we had already such cases during the development

[ben_hamburg]

Can you remember any specifics on the case where you had this bug?
Maybe I can pin down whats causing it, if it´s the same issue.
The particle flow contains a lock/bond OP, maybe that is causing trouble?

[ben_hamburg]

Hm, now I also noticed something else that seems weird:

Shouldn´t the particle count increase with increasing grid resolution? With liquid simulatons the foam count increases with increasing grid resolution, but with my drag particles it seems to be the other way around:
The more I increase the grid resolution, the less particles I can see...
But when I lower the grid resolution, the grid is visible in the particle formations.
I´ll try to find some time tomorrow to do all these tests with my empty test scene , just checking in to see if there is a bug or the scene is messed up, or if I´m missing something...

[Ivaylo Katev]

the particles are created by source, not born by the simulation, so their count must remain when you increase the grid size.
i can't see the particle count in the image, but definitively it looks low, about few thousand. the source value is really high, so i suppose the particles are just produced in the same positions over and over again. try to use inject move instead brush, it pushes the particles out and must produce some dynamics that will scramble them.

[ben_hamburg]

Well, I increased the particle count to 100 mill and now it looks good enough for me, So i´ll keep the current sim for now.
I´ll run another test with inject mode and see if that changes anything, if not I´ll send a scene to support for investigation.
It´s still amazing that Phoenix can simulate 100 mill particles without stuttering though...

by the way: Can I somewhere see the number of particles actually beeing simulated? With foam I can see the foam particle count in the Simulation tab afaik, but i can´t see the particle count there from my drag particles.

[Ivaylo Katev]

well, i noticed that you use the particle birth parameter as particle count
phoenix really can handle 100M particles, but one frame takes about 30-40 sec, and the cache size is about 1G per frame
the actual particle count is placed in the main simulation panel, there is info about the loaded frame, including the particles in it.