Liquid in a cyclone

[JuhaHo]

Actually, it seems that Path Follow influence has no effect if it's other than 0. It's kind of on-off switch. Are others being able to confirm this?

[georgi.zhekov]

Seems to be working correctly on our side - can you send over the scene that is not working so that we can take a look?

[JuhaHo]

At the moment I'm investigating if it's tyFlow or Phoenix. It's not an easy task.

[fixeighted]

It seems fairly simple to get a result like the attached, which appears to do what it is supposed to, based on reference I have seen, but
of course would need refining. This just uses a path follow and some tweaked parameters of that and the grid and liquid settings.

I'm not convinced that PHD can do such a specific sim based on the necessary real world physics of this apparatus, given that that so many
parameters need to be assessed. I'd love it if the devs could confirm that, or it's just my lack of tech knowledge
This is based on looking at various videos using Solidworks or other software for the simulation. It's a pity they don't
export to any usable format for Max.

[JuhaHo]

Thanks,

That's the point where I'm able to go with Phoenix too. The problems start, when the cyclone is filled and it should start to do the job, it has been designed for. When the system is filled in Phoenix simulation, the liquid don't travel down anymore. Instead it turns immediately to outlet and the liquid leaves the system uncleaned. It kind of loses it's inertial energy and stops moving down forcing the liquid ahead to go up.

[fixeighted]

Yes, that is what I mean by not being convinced that Phoenix can accurately represent all the very specific behaviours that this needs.
Therefore you have to fake it. Those other simulators are designed specifically to represent real behaviour to a very high degree,whilst PHD is more of an artistic tool to
get you a long way there but without the absolute accuracy.

Again, maybe Georgi can elaborate on whether he believes the same can be done in PHD as in Solidworks for example.

Split it into a few parts and do it that way would be my approach. Your client can't expect miracles when there are none to go around

[JuhaHo]

I might have one workaround. I just forget the Path Follow method as it has not been designed to be used with these kind of situations. I would need to do two Phoenix simulations: one for rendering and one for tyFlow particle simulation. Since Phoenix can't handle the real life situation, where the valves on the bottom are operating, I'll just leave them open for tyFlow version of Phoenix simulation. Then I do another Phoenix simulation with animated valves for rendering purposes. On that simulation, the fluid voxels are trembling on spot, because that's how Phoenix currently works. The illusion of movement comes from tyFlow particles made in open valve run so the render material for Phoenix scene objects has to be homogenous.

In tyFlow the particles must be sent to another event before the particles reach the valves. That has to be without Fluid Force operator. This way the particles are not flushed away due to non non animated valves in the first Phoenix simulation.

I'm out of office right now and I can't allocate PhoenixFD license to my mobile workstation so I can't test this new approach right now. I'll try this approach when I get back to office.

[JuhaHo]

My first simulation with this new approach didn't quite work as I wished. Everything goes down through opened bottom valve because the cyclone-effect does not work in Phoenix physics. I think next I try opening the valve half way.

[JuhaHo]

So I made the lower valve just a tiny bit open. Now I got the continuous down flow. When I'm satisfied with the particles, I'll do another run with working valves for the fluid part. Anyway, it's the client who tell if I got it right.

[JuhaHo]

I'm now finetuning the particle amount at the junk trap before scheduled flushing. Almost there. If this get's accepted, there's more and different cyclones to go.