Page 1 of 1

K_Volume Problem

Posted: Wed Aug 11, 2010 2:16 am
by BrianLooney
Hello RF Team,

I have found a situation that may or may not be a bug.

Here is the situation:
I have a mesh object to hold a grid fluid.
I filled up the mesh object with grid fluid particles.
Image

Next I want to lower the particle amount (I over-filled it),
so I add a k-volume daemon,
check the inverse option, and simulate a frame.
Image

Now I remove the k-volume daemon,
but the grid fluid keeps deleting particles as if the daemon was still present.
I tried turning the k-volume inactive, removing it from the global links,
and even deleted it completely from the scene,
but the grid fluid still acts like the daemon is still there for all future simulation.
See below, the daemon is gone,
but the grid fluid still gets killed as if the daemon were still in the scene.
This was after several more simulated frames after the daemon was deleted:
Image

Re: K_Volume Problem

Posted: Wed Aug 11, 2010 11:10 am
by Alex
Hi Brian,

we have been looking into this and it is definitively a bug.

This is what happens:
1. A kill volume sets which cells must kill particles. It remains that way forever unless...
2. A kill volume is moving and intersects at some place with the grid domain bounding box. In that case "killer cells" are computed every time the daemon changed its shape.

You can also reproduce some strange behaviours when using several K_Volumes if not all of them are moving.

So, how to reset the cells to avoid your behaviour?
When you are ok with the hole the kill volume has made, set it as inactive and place another kill volume daemon at some location within the domain but without fluid. Add some little animation to this last daemon.

Thanks for finding this bug. Please let us know if the workaround helps.

Re: K_Volume Problem

Posted: Wed Aug 11, 2010 1:10 pm
by BrianLooney
Thanks for the response, and for the work-around.

Re: K_Volume Problem

Posted: Sun Nov 28, 2010 1:49 pm
by bitseq
Has this bug been fixed yet? I also seem to be having problems, I can't seem to get th k_deamon to work on my grid particles, should the deamon not intersect the grid domain boundary in order to work correctly?

Re: K_Volume Problem

Posted: Mon Nov 29, 2010 11:46 am
by LuisMiguel
Hi bitseq,

it seems has not been fixed yet. We are working in it. Sorry for the inconveniences.

luisM.

Re: K_Volume Problem

Posted: Mon Nov 29, 2010 2:08 pm
by bitseq
OK Thanks