Liquid in a cyclone

[JuhaHo]

Thanks,

I tested this approach. It seems not to be working. As soon as the influence parameter is increased from zero, it kind of sucks the liquid out of the junk trap leaving it empty. I need to figure out something else.

[JuhaHo]

I'm starting to get some results after iterating the Max Distance parameter. So far the best value has been 400 mm (if it uses scene units). Now I will iterate the influence parameter.

[JuhaHo]

Here's another try. I think there's still a lot of tweaking to do. I think, I need to focus on iterating both: the Max distance and influence value, the latter being animated.

[fixeighted]

Could you not simply do what mainly you've done here, though continue the path so it has a central tight helix which the liquid can travel back up on and out of the outlet?
This is what I tried out and it works visually plausibly and with little effort.

[JuhaHo]

Here's a new screenshot, I made with tyFlow simulation. I think, this clearly shows what is happening inside Phoenix simulation. The spiral should not be visible. Yes, the fluid goes down on spiral path, but it goes much more evenly distributed along the cyclone walls. Is it the pipe radius, I should start playing with? I currently have the default 5 mm, while the cyclone radius is about 350 mm

fixeighted I have tested plenty of paths. On my screenshot, there wasn't the right one. The path used for this includes the up flow in the middle.

[fixeighted]

If this is for illustrative purposes, i.e. so that the client/viewer can clearly identify what's going on, then I'd suggest not going for absolutely correct physical behaviour, as that may be very hard to read
and, judging by the length of this post, may not reach a satisfactory endpoint.

So why not get the main indicatative spiralling going as it is it now, then add either a secondary sim to coat the entire chamber sides or
a geometry method using animated displacement.
If that is thin enough then you'll should get a result which visually describes the process.

Just suggestions, so take with appropriate amounts of sodium

[JuhaHo]

Thanks,

It's the client, who says when this is good to go. I've already spent one year daily working to get this acceptable, but so far I'm not even near getting this done. If all goes well, I would expect another 2 - 3 years to get this fixed, if I don't run out of money doing this.

The client has a professional fluid simulator and I have lots of flowline screenshots on how the voxels are travelling through the cyclone. I'm trying to replicate the results with Chaos PhoenixFD, but I can't get nothing similar to that. Here's my yesterdays results and I'm clearly going from bad to worse.

[JuhaHo]

I really don't know how to continue. I think Path Follow helper is not suitable for a cyclone animation. I just don't work in this context, Is there anyone in chaos, who can tell how to use basic fluid physics in PhoenixFD liquid animation?

Please.

[georgi.zhekov]

Hey,

Have you tried a vortex force and something to pull the liquid down a bit?

[JuhaHo]

Thanks,

I'm no sure how to proceed. I tested Vortex, but didn't notice any difference on the results. Maybe I did it in a wrong way.

I think I need an upward force in the core. It would make room for the voxels on the outer ring to go down first.

[JuhaHo]

I really don't know how to proceed. Every suggestion on how to get this working would be appreciated.

What do I need to do keep the fluid moving? So far my attempt has been disastrous.

[georgi.zhekov]

Can you show us what happens currently in motion? What do you mean by keep moving - is the liquid stopping at some point? You want the liquid to spread more around the walls?

[JuhaHo]

Here's the difference. Path Follow has just been activated and it's tarting to create a spiral from top to bottom. I want the conservation to be more or less uniform.

[georgi.zhekov]

You want it to be more like the bottom part and not so like it is in a tube like the top part where the X is right?

[JuhaHo]

Yes,

However Here's a screenshot of the last night simulation run flowlines. With these flowlines I would expect more even distribution, but no.