Strange box around object hybrido ...

[ValentinD]

Hello FlorianK,

I am using 5 - 50 sub-steps.

Thanks for the interest !

[FlorianK]

Hybrido Solver stepping or Hybrido Secondaries? I am also not interested in the General Tab. Also, are you using any Vorticity Boost?

[ValentinD]

Ok, here is my parameters :

Simulation Options :
general :
Min : 1
Max : 300

----
HY main :
Min sub : 5
Max Sub : 50
Strictness 0.5

Pressure Solver :
Max i : 1000
Accuracy : .375

Transport : Particle steps : 7

----
Hy secondary :
Min : 1
Max : 80
(I have no splash in my scene for now)

----
OpenCl - GPU : checked

---
Scale options :
Global GeoScale : 0.05
CEll size scale : 0.8
Deamon force scale : 20

---
HY_Domain :
Cell size : 0.2
No vorticity boost on my domain;

---
All Objects : Every bamboo stick of the jail
Cell size : 0.2
Surface Offset : 0.15
Friction bandwidth : 1
Interaction factor : 1


Thanks a lot FlorianK

[FlorianK]

Ok, here is my parameters :

---
Scale options :

Global GeoScale : 0.05

I guess that's where the core of the problem is. Recreate your scene with a scale setup of 1.0 inside Realflow. Scale your geometry inside your 3D App prior to exporting so it has a scale one can easily work with inside RF. I bet, then you can simulate with 1-3 Substeps for Hybrido and will have no problmes with the jittering.

[ValentinD]

I will try it, I don't understand why the geo scale create so many problems... That's not the first time that I had an issue with...

My scene is a bit gigantic in realflow (1 maya unit = 1 cm) so i will try it...

I will give you some news as soon as possible !

Thanks for the help !

[ValentinD]

Hello FlorianK !

It seems that the global scale option is in charge of all this trouble ! I have a better simulation now !

Did you have any advice for some meshing parameters ? I think that I will update this thread for the meshing part ...

[FlorianK]

The issues with tiny scales aren't something to be found only inside Realflow. You'll experience similar problems with other solvers in other packages. A huge amount of interdependent values has to be considered when scaling things down that much, it is not only the geometry's size. For me, it works best and fastest working in a 1.0 scale inside RF. For the mashing...what's your issue with that?

[ValentinD]

You are right it is much faster and accurate with scale 1... I will remember it, thanks a lot for that !

I had a small issue with the meshing... Lonely particules are pretty big when meshed and under tesselate...

Here is a screenshot :

https://mega.co.nz/#!gVw31aDD!DjZe6ib7N ... fZzivlBRvE

Thanks a lot !