Hello,
I would like to know if it’s possible to define bond forces between two different materials
(for example copper and diamond for an indentation).
If yes, how to do it in LAMMPS?
Thank you in advance,
Hugo
Hello,
I would like to know if it's possible to define bond forces between two
different materials
(for example copper and diamond for an indentation).
If yes, how to do it in LAMMPS?
this is the wrong question to ask at this point. before going this
far, you should ask *yourself* two other questions:
1) what do you want to learn from your simulation?
as diamond is quite hard and copper pretty soft by comparison. could
it be sufficient to use an idealized indenter as it is used in the
peridynamics example provided in the LAMMPS distribution?
2) is there a theory/model for this?
would a mesoscopic model (as peridynamics is) even be able to
represent this or do you have to use atomic resolution and suitable
atomic interactions? this is something that a thorough literature
search could answer better than a mailing list.
now, if you actually reach the conclusion that you still want to do
peridynamics, you have to check whether the model you need to use, is
available in LAMMPS and whether you have suitable input parameters for
it. if you still have a problem at that point, let the list know.
axel.
The answer to the PD part of your Q could
come from the pair peri doc page - it lists
the PD models that LAMMPS supports
and their inputs. It looks like bond strength
is set by s00 and alpha. And just one set
of those values is specified. Rezwan can
comment on whether it could be allowed to
have different bond strenghts per material
and whether that makes sense.
Steve
You can model the material interface using peridynamics. For this you need to define the effective elastic modulii, strength, s00, etc for the interface between two materials. The thesis PERIDYNAMIC THEORY FOR MODELING THREE-DIMENSIONAL DAMAGE GROWTH IN METALLIC AND COMPOSITE STRUCTURES by Erkan Oterkus (also check the presentation slides: http://www.bristol.ac.uk/composites/events/2012/oterkus-presentation.pdf) has some good information on this. For composites the interaction between matrix and fiber is defined by this effective pari_peri interaction. In this thesis bond based PD is used for this problem. Also you can look the recent Sandia report "Origin and Effect of Nonlocality in a Composite" by Stewart A. Silling for modeling nonlocality in the composites.
If you want to consider the length scale as one of the parameters you can link the atomistic scale information on materials interface to the coarser length scale models. Recently, a peridynamics based hierarchical multiscale modeling scheme (PFHMM) is developed. This can be applied/extended for your problem. PFHMM used PDLAMMPS and LAMMPS in order to pass information back and forth. See the following papers:
Rahman, Rezwanur, John T. Foster, and Anwarul Haque. "A multiscale modeling scheme based on peridynamic theory." International Journal for Multiscale Computational Engineering 12.3 (2014).
Rahman, R., and J. T. Foster. "Bridging the length scales through nonlocal hierarchical multiscale modeling scheme." Computational Materials Science 92 (2014): 401-415.