Simulating particles & liquid in a cyclone

[JuhaHo]

Has anyone ever tried to simulate cyclone physics. I'm getting some results, but I have quite a few issues too. The liquid infeed is on the lower and the outfeed is on top of the cyclone. The liquid contains some stones, which should drop at the bottom. The particles leave the spiral path too early and drop when they should still be inside the cyclone. The particles are driven by liquid, so maybe it's a PhoenixFD problem.

[JuhaHo]

I have had this issue with other animations too. I wonder, what I should do to keep the particles inside the liquid. As far as I know, the force affecting to particles are generated by the Phoenid liquid.

[JuhaHo]

Any ideas how I can distribute the particles more evenly inside the cyclone?

[betamax]

Any ideas how I can distribute the particles more evenly inside the cyclone?

Depends where they are being born from? Maybe birth the particles in the entire volume of the cyclone?
Might help to see a screengrab of your tyflow.

[d4rk3lf]

I might not understand question... but...

Can't you simply create helix, turn on "enable in viewport", so you get it as mesh object, animate helix spinning, then use Tyflow surface force with helix to attract particles to it?
(or maybe even path find target, instead of surface force)

[JuhaHo]

Thanks,

The only force apart from gravity is Phoenix FD. The Helix path controls Phoenix and the particles are born at the same place where the Phoenix Emitter is. The path follow in tyFlow has been deactivated. The liquid is picking up the speed when going down and starts rising in the middle of the cyclone. The particles being heavy continue down eventually ending to particle trap under the cyclone.

[tyFlow]

It's hard to say what the cause is from just single screenshots...or even what the problem is. Are you trying to get the particles to follow the helix more closely? Or be more spread out while still following the overall path of the helix? Is this meant to be a realistic sim or more illustrative?

[JuhaHo]

It's hard to say what the cause is from just single screenshots...or even what the problem is. Are you trying to get the particles to follow the helix more closely? Or be more spread out while still following the overall path of the helix? Is this meant to be a realistic sim or more illustrative?

Thanks,

On the contrary, the particles are supposed the follow the liquid all the way down. The helix is for Phoenix FD path force. The problem is that particles are following the Helix path force while they should be evenly distributed in liquid. Of course if Phoenix behaved like in real physics, we would not need additional path force. I have seen advanced simulation results made with industrial simulators, but I don't know if the vectors can be used here. Those results are quite different to Phoenix results.

I'm trying to get some level of realism to the fluid/particle behavior, how ever in rendering realism is not needed, because we want to show what is happening inside the cyclone.

[tyFlow]

If you want particles to follow a fluid sim, the Fluid Force is the operator you want to start with.

Clear all your other PhysX/Speed/etc operators out, and start with Fluid Force....then work your way forward with collisions and whatever else you need.

[JuhaHo]

I thought PhysX fluid is needed for this kind of two-way simulation. Anyway, I modified the geometry to decrease the counter pressure at the bottom. Now I hope, most of the fluid will go up in the middle while the particles continue fall down. The conical shape in the middle is a collision geometry for shapes only, so they wouldn't follow the fluid upwards.

[tyFlow]

For fine-tuned animation like this - especially when you're having trouble getting the very basic motion working - you're not going to want to go with a 2-way solution, which is going to be difficult to control. If you really need to see interactions between your fluid and the particles, I would suggest doing a starting fluid simulation, then advecting particles through it with the Fluid Force operator, then doing a secondary fluid simulation using the particles as colliders. The particles will be moving with the overall (primary) fluid giving the illusion of 2-way interactions, even though interactions with the secondary fluid sim will only be 1-way.

[JuhaHo]

Thanks,

So I create the Phoenix simulation tyFlow disabled and after that I do the tyFlow simulation where I replace the PhysX fluid with FluidForce operator?

[tyFlow]

Yea, that would be a good starting point I think.

[JuhaHo]

I have now finished first tests by using the fluid force instead of PhysX fluid. I also replaced PhysX collision with collision. This caused a new issue; small portion of particles follow the liquid through one collider. This didn't happen with PhysX collision.

Can Fluid force and PhysX collision be used simultaneously?

[JuhaHo]

I have made quite a few simulation runs recently. There's one thing that ruins things. No matter what I do the particles are escaping the container. They follow the liquid, but outside the container. Is there anything I should know. I want the particles to stay inside the fluid which is inside the container.

[tyFlow]

Have you set the container as a collider (using a PhysX Collision op if your particles are PhysX particles)?

Alternatively you can use a Surface Test to do an inside/outside test and delete any rogue particles that escape.

[JuhaHo]

Have you set the container as a collider (using a PhysX Collision op if your particles are PhysX particles)?

Alternatively you can use a Surface Test to do an inside/outside test and delete any rogue particles that escape.

Thanks,

I was suggested earlier to give up PhysX and use tyFlow operators instead. So I'm using tyFlow Collision, Force (gravity) and Fluid Force.

[JuhaHo]

So, how should I tell the particles, that they should stay inside the container, not fly around? I know how and where the particles should be traveling, but I don't know how to tell that to tyFlow.

[tyFlow]

You can prevent particles from leaving the container with several operators, depending on your needs. For example, you can use a Collision operator to have the particles bounce off the container. You could use a Surface Test to test for particles that leave the container and delete them. Or you could use a Push In/Out operator to just push any outside particles back in. Make sure the timing of whichever operator you choose is set to continuous.

[JuhaHo]

Thanks,

I'm now using Collosion objects and the situation is that not a single particle of tens of thousands finds it's way to the final destination. Would it be better to use PhysX particles&collision?

Earlier when using PFlow&Orbaz tools, I baked the particles into individual meshes to separate scene file. Then I opened the scene and deleted the particles manually there. This produced very large files, that made the scenes non renderable due to backburner timeouts. Of course I would now be able to read the particles into tyCache.

[tyFlow]

Well it's all dependent on your setup, goals, etc. Each method has its own caveats, benefits, etc.

I can't guess why your particles aren't getting to their destination. You have to show your setup.

[JuhaHo]

Here's a screenshot of my system. I hope this tells something. It seems PhysX collision is much better than Collision. There's also timing, where I want the particles stop following the liquid.

[tyFlow]

Yes, if you are using PhysX Shapes, you need to use a PhysX Collision operator.

[JuhaHo]

Thanks

I didn't use PhysX shapes with the non PhysX collision. I activated those for PhysX collision.

[JuhaHo]

I still don't get it. Any ideas what I'm doing wrong here.

1. The particles go trough bottom valve, while the liquid stays there.
2. Particles inside the cyclone are in tight chain around the Phoenix path force splines.