balls initially touching walls

Hi,

I am creating spheres so that they initially intersect with walls. I get the following message

WARNING: Dangerous build in triangle neighbor list.

What is the consequence of this? I am getting a memory error I don't understand and I wonder if this is the reason...

Thanks for all the help,
George

Forums:

LIGGGHTS® - User Forum

raguelmoon | Tue, 11/02/2010 - 06:28

Hi George, Please read the

Hi George,
Please read the doc pages for this kind of warnings:

"When using style 'mesh/gran' in combination with the 'fix pour' command, you always have to keep a minimum distance between the wall and the insertion region that is equal to maximum particle radius + half the neighbor skin defined in the neighbor command. "

Best,
Ram

marketos | Tue, 11/02/2010 - 12:55

Hi guys, Thanks for the

Hi guys,

Thanks for the replies.
If possible I would like to be able to take a sample created in a different programme which is already carrying force and plug it in into LIGGGHTS. So I think it should be generally possible to create something that runs fine in LIGGGHTS even with an initial overlap. But I am not sure if this is what causes my memory-related error.
When I asked what the consequence is, I meant, does the warning appear because the neighbour list is not generated well? (I agree that energy is created.) The reason why I was asking was that I have a slightly weird situation where my sample box made from meshgran walls is not very wide so that a sphere can possibly touch with a large number of wall triangles- maybe this is the cause of my memory error?

Thanks,

George

ckloss | Tue, 11/02/2010 - 13:37

The warning appears b/c there

The warning appears b/c there is an initial overlap, but it does not say if this is severe or not.

>>When I asked what the consequence is, I meant, does the warning appear because the neighbour list is not generated well?
This could be the case - if the overlap is small, the simulation should still be valid.
As I said, you can try using the fix nve/limit to do some relaxation and after say a couple k steps, change to nve/sphere
If the memory error persists, please send me the backtrace

Cheers,
Christoph

ckloss | Tue, 11/02/2010 - 09:30

>>I am creating spheres so

>>I am creating spheres so that they initially intersect with walls
What is the intention to do this?

>>What is the consequence of this?
The consequence is that overlap energy is created which is in general not what you want. Especially, it can blow your simulation if overlaps are too large. What you can do is to use fix nve/limit for a few relaxation steps.

>>I am getting a memory error
Do you get the error also if the particles do not initially intersect the wall? if yes - can you post that please. thanks