Restitution coefficient of rigid bodys

Submitted by Nafets on Tue, 11/04/2014 - 09:37

Hi all,
I generate a rigid body shaped like a box with moltemplate and simulate a droptest. The problem is that the body jump's after contact with the mesh to a height that is not in relation to the physical law. k = square root (h'/h)
Restitution coefficient: 0,5
Previous height: 0,5 m
Height after first contact: 0,27756295 m
According to the formula the body should just jump to a height of 0,125m.

Forums:

ZYan | Tue, 11/04/2014 - 09:51

Hi Nafets,
Did you use fix rigid command to create the rigid body? and did you use public version liggghts? if both yes, then it is not strange you got this results as the 'overlaped' mass in the body is ignored.
Cheers,
Zyan

Nafets | Tue, 11/04/2014 - 09:57

Hi ZYan,
Yes i use fix rigid and the public version of liggghts. Ok so this means that with the public version it is not possible to get realistic output with the real restitution coefficient.
Thanks for your answer

ZYan | Tue, 11/04/2014 - 10:04

Hi Nafets,
I would rather say the restitution coefficient is correct but the mass of the body and the moment of inertia tensor are not correct.
Chees,
Zyan

ckloss's picture

ckloss | Wed, 11/05/2014 - 19:56

>>I generate a rigid body shaped like a box with moltemplate and simulate a droptest.
>>The problem is that the body jump's after contact with the mesh to a height that is
>>not in relation to the physical law. k = square root (h'/h)

Correct! Fix rigid is not capable of correctly recovering the coefficient of restitution!
There's an article on this effect:http://www.sciencedirect.com/science/article/pii/S0032591008002143
We have a multisphere implementation in our commerical support version which remedies this.
We'll communicate that more clearly in the near future.

Best wishes
Christoph

ma | Fri, 06/19/2015 - 10:43

Hi,
Is the problem with the restitution coefficient solved in the latest version?
I am modeling rigid bodies that fall out of the boundaries during simulation.
thank you
ma

ma | Fri, 06/19/2015 - 10:48

In addition, I am using E=1e8, that leads to a realistic bahavior in the case of single spherical particles model and I am going to compare the results. So, I can not reduce E
thank you
ma

ckloss's picture

ckloss | Wed, 07/01/2015 - 14:27

Hi,

We have a multisphere implementation in our commerical support (premium) support version which remedies this.

best wishes
Christoph

ma | Tue, 07/14/2015 - 16:40

Hi,
It means that I cannot model it with public version?