Problem with Sheeter Daemon

Describe your realflow bugs
duke3d
Posts: 22
Joined: Tue Aug 30, 2016 11:54 pm

Problem with Sheeter Daemon

Postby duke3d » Fri Sep 23, 2016 9:35 am

here is what I got https://www.youtube.com/watch?v=EZoNZNLs7-8

In general Sheeter Daemon can't "see" the wall of a glass. It keeps generating particles on the outer side (no particles inside wall) until particles in the glass stop moving. I tried different Sheeter settings, different Cell Size and played with Surface Offset, nothing helped. If I switch off Sheeter Daemon, no problem occurs, but I can't get desire behavior of a liquid :(

Anyone had the similar problem?
Any suggestion?

Thanks,
Andrew


User avatar
ChristianZ
Posts: 3257
Joined: Sat Sep 09, 2006 11:37 am
Location: CH

Re: Problem with Sheeter Daemon

Postby ChristianZ » Fri Sep 23, 2016 2:05 pm

Hi

I see, that could be tricky. You need a method to either stopp the particles at the border of the glass or destroy the escaping particles.
There are methods to destroy particles inside a certain volume, but otherwise it might be hard. Could you disable the sheeter for the time that you pour in the liquid and turn it on only later?
RF10 standard, System specs: OSX 10.11, MacPro 4.1 2xXeon, GTX980Ti, 64 GB RAM

User avatar
ChristianZ
Posts: 3257
Joined: Sat Sep 09, 2006 11:37 am
Location: CH

Re: Problem with Sheeter Daemon

Postby ChristianZ » Fri Sep 23, 2016 6:26 pm

I made a little test with a collision situation of dyverso particles without and with sheeter daemon. That went well in this example.
https://youtu.be/EJyka1Bcz50
scene file for RF2015 standard
http://www.cardiomyocytes.com/realflow/ ... lision.zip

Of course, this is a relatively simple piece of geometry here with a thick wall and the object mode is "solid inside". So, it may be possible to improve the geometry of the glass or use a thicker collision object just for the purpose of the simulation.
RF10 standard, System specs: OSX 10.11, MacPro 4.1 2xXeon, GTX980Ti, 64 GB RAM

duke3d
Posts: 22
Joined: Tue Aug 30, 2016 11:54 pm

Re: Problem with Sheeter Daemon

Postby duke3d » Sun Sep 25, 2016 6:22 am

well, the problem is not about particles escaping glass or other cans ;) It's about Sheeter that connects the particles which are outside the glass and those that are inside. It doesn't see any barrier (a wall of a glass) between them. Even if I increase the cell size, Sheeter keeps generating particles, though the distance between particles in the glass and outside the glass is huge. I've made a new simulation with Sheeter Daemon on and off https://youtu.be/uMQT6_UldWM

User avatar
ChristianZ
Posts: 3257
Joined: Sat Sep 09, 2006 11:37 am
Location: CH

Re: Problem with Sheeter Daemon

Postby ChristianZ » Sun Sep 25, 2016 6:59 am

Sure, the sheeter is filling holes and it does not care if there is something else between the particles. This is how the sheeter works, but then there is the parameter "Min cavity size". This is the threshold for the size of a hole, larger holes than that will not be filled. For your scene this means to reduce it, to a value of 0.5 for example.
If there are too many new particles made in this way, you can still reduce the "detection ratio" in order to avoid that.
RF10 standard, System specs: OSX 10.11, MacPro 4.1 2xXeon, GTX980Ti, 64 GB RAM

duke3d
Posts: 22
Joined: Tue Aug 30, 2016 11:54 pm

Re: Problem with Sheeter Daemon

Postby duke3d » Sun Sep 25, 2016 9:06 am

yeah you are right, that is how it should be. When I reduce "Min cavity size" to anything between 0.0 and 0.9999 Sheeter fills no holes at all O.o

User avatar
ChristianZ
Posts: 3257
Joined: Sat Sep 09, 2006 11:37 am
Location: CH

Re: Problem with Sheeter Daemon

Postby ChristianZ » Sun Sep 25, 2016 10:12 am

If you have not yet read it I can recommend the latest issue of the waterline tutorials. http://www.waterline.tv/
There is a very detailed discussion about the sheeter and scripts inside.
RF10 standard, System specs: OSX 10.11, MacPro 4.1 2xXeon, GTX980Ti, 64 GB RAM

duke3d
Posts: 22
Joined: Tue Aug 30, 2016 11:54 pm

Re: Problem with Sheeter Daemon

Postby duke3d » Mon Sep 26, 2016 11:14 am

duke3d wrote:yeah you are right, that is how it should be. When I reduce "Min cavity size" to anything between 0.0 and 0.9999 Sheeter fills no holes at all O.o


oh my bad :))) I meant "Cavities detection ratio" ))) Anyway, thanks for the link :D By the way I've read "Splahes And Tendrils" about a months ago, it's was quiet helpful :) However, I've tried "k-Volume graph" form "RealFlow Fields And Graphs", It didn't help :(

User avatar
ChristianZ
Posts: 3257
Joined: Sat Sep 09, 2006 11:37 am
Location: CH

Re: Problem with Sheeter Daemon

Postby ChristianZ » Mon Sep 26, 2016 11:24 am

But it almost works, right? I think, the best would be to make the sheeter less active by reducing the cavity setting, After all, this classic scene of filling a glass rather needs a high surface tension to appear more realistic, unless you want to make a scene with a completely different type of fluid than water.
RF10 standard, System specs: OSX 10.11, MacPro 4.1 2xXeon, GTX980Ti, 64 GB RAM

duke3d
Posts: 22
Joined: Tue Aug 30, 2016 11:54 pm

Re: Problem with Sheeter Daemon

Postby duke3d » Mon Sep 26, 2016 12:05 pm

ChristianZ wrote:But it almost works, right? I think, the best would be to make the sheeter less active by reducing the cavity setting, After all, this classic scene of filling a glass rather needs a high surface tension to appear more realistic, unless you want to make a scene with a completely different type of fluid than water.


well you are right. I'm working on orange juice. I was able to get right behavior of liquid very fast. The problem occurred when I wanted to make splashes and juice flowing down on outer wall of a glass. So, I got that strange particle generation by Sheeter on the outer wall.
Reducing "cavity detection ratio" helps just a bit. Reducing it to much ruins nice liquid behavior. Sheeter is awesome tool, but if someone has liquid to be simulated inside and outside of a glass (cup, can, box, etc) at the same time, he/she will spend huge amount of time for simulation to get right result.


Return to “Realflow bug report”

Who is online

Users browsing this forum: No registered users and 1 guest