[Openmd-users] Custom rigidBody and defining force field
Dan Gezelter
gezelter at nd.edu
Mon Dec 1 10:47:45 EST 2014
Martin,
It looks like your icosahedra are overlapping - the Hmat line in your Ico.md file defines a box that is only 20 Angstroms on a side. In this periodic box, x coordinates of +10 and -10 are actually the same location. You can modify the Hmat parameters to define a larger box, or if you don't want periodic boundary conditions, you can just add the parameter:
usePeriodicBoundaryConditions="false";
which will make OpenMD ignore the Hmat line entirely.
One hint on Dump2XYZ, using the -m flag will map all coordinates that have gone outside the box back into the box on the other side. This flag can be quite helpful for diagnosing box size problems.
Best regards,
--Dan
> On Dec 1, 2014, at 10:39 AM, Martin Vala <mvala at saske.sk> wrote:
>
> Hi Dan,
>
> I just found out that 'x' position of both rigid bodies (10 and -10) are big for some reason. When i set it to 3.0 and -3.0, simulation works
>
> Can you give me hint, what am i doing wrong?
>
> Ciao
>
> Martin
>
> On 01.12.2014 00:15, Martin Vala wrote:
>> Hi Dan,
>>
>> Thanks for information it is working nicely.
>>
>> Now i wanted to try to put 2 ico rigid bodies next to each other. But i am getting following error while simulation + Error when i am converting it to XYZ via Dump2XYZ
>>
>> Can you give me hint? Files are attached
>>
>>
>> OpenMD info:
>> ForceManager::setupCutoffs: No value was set for the cutoffPolicy.
>> OpenMD will use TRADITIONAL.
>>
>> OpenMD info:
>> StatWriter detected a numerical error writing: Total Energy
>> Converting Dump to XYZ ...
>> OpenMD FATAL ERROR:
>> DumpReader: Ico.dump does not contain a valid frame
>>
>>
>> Ciao
>>
>> Martin
>>
>> On 27.11.2014 20:46, Dan Gezelter wrote:
>>>> 1. How can i add other atoms to interact with this Ico rigid body. And what do i need to add to Martin.frc file to setup interaction. Is it enough to setup forces? Like it is explained in openmd manual.
>>>>
>>> Yes, you just need to make entries for the new atom types in the frc file and to provide some interaction model between the atoms on the RB and the other atoms. You'll also need a molecule definition, a component, and entries in the <Snapshot>. See the new demo files attached below.
>>>
>>>
>>>> 2. How can i make my rigid body fixed in positions (no movements) and when it will interact with other (new atoms), it will not move.
>>>>
>>> You could make a few (or all) of the atoms in the rigid body super heavy (i.e. change their mass to 10000 or more). We sometimes do this for nanoparticle simulations to prevent brownian motion of the particles away from a fixed location. You'll need more than one heavy atom if you want it to stop rotating as well. You could also add positional and orientational restraints, but these are harmonic restoring forces, so they will allow some motion.
>>>
>>> Best regards,
>>>
>>> --Dan
>>>
>>>
>>>
>>>
>>>
>>>> Thanks again
>>>>
>>>> Ciao
>>>>
>>>> Martin
>>>>
>>>>
>>>> On 27.11.2014 20:05, Dan Gezelter wrote:
>>>>
>>>>> Martin,
>>>>>
>>>>> You do need to define the atom type in a force field file, and the force field can be quite a bit simpler than Amber for testing. The other issue I saw was that the quaternions in this line:
>>>>>
>>>>> 0 pvqj 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 -1.8e-05 1.6e-06 1.9e-05
>>>>>
>>>>> only had 3 entries. There should be 3 positions, 3 velocities, 4 quaternions, and 3 angular momentum variables. The convention we use for quaternions is defined on page 88 of Allen & Tildesley's "Computer Simulation of Liquids" book. To orient the principal body axis along the z-spatial coordinate, we could choose the following Euler angles:
>>>>>
>>>>> \theta = 0
>>>>> \phi = 0
>>>>> \psi = 0
>>>>>
>>>>> These correspond to a set of quaternions q = (1, 0, 0, 0).
>>>>>
>>>>> Anyway, I tested the three attached files, and you should be able to experiment further with these. A couple of changes I made: 1) The _ (underscore) character has special meaning in the selection syntax, so I modified your atom types and rigid body definition. 2) The angular momentum was quite small, so I boosted it a bit so that you can observe the rotation in this simulation.
>>>>>
>>>>> Try this with the attached files:
>>>>>
>>>>> openmd Ico.md
>>>>> Dump2XYZ -i Ico.dump
>>>>> jmol Ico.xyz then go to "Tools" -> "Animate" -> "Once"
>>>>>
>>>>> Best regards,
>>>>>
>>>>> --Dan Gezelter
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>> On Nov 26, 2014, at 11:01 PM, Martin Vala <mvala at saske.sk>
>>>>>>
>>>>>> wrote:
>>>>>>
>>>>>> Hi all,
>>>>>>
>>>>>> i just notice that i didn't define rigid body in my previous configuration. I am attaching my updated config files. I am sorry for wrong configuration files.
>>>>>>
>>>>>> Ciao
>>>>>>
>>>>>> Martin
>>>>>>
>>>>>> On 27.11.2014 11:37, Martin Vala wrote:
>>>>>>
>>>>>>
>>>>>>> Hi All,
>>>>>>>
>>>>>>> sorry i misstyped my rigid body. It is icosahedron
>>>>>>>
>>>>>>> Ciao
>>>>>>>
>>>>>>> Martin
>>>>>>>
>>>>>>> On 27.11.2014 11:12, Martin Vala wrote:
>>>>>>>
>>>>>>>
>>>>>>>> Dear OpenMD users,
>>>>>>>>
>>>>>>>> I would like to define my own rigid body with my own atoms (my custom atom is "My"). My rigid body is defined as icohexadral (in RB_My.md file - attached). I am also attaching RigidBody_01.md file.
>>>>>>>>
>>>>>>>> Problem:
>>>>>>>>
>>>>>>>> OpenMD FATAL ERROR:
>>>>>>>> Can not find Matching Atom Type for[My_RB]
>>>>>>>>
>>>>>>>> I am guessing that "My_RB" atom type is not defined in $FORCE_PARAM_PATH/Amber.frc (in my RB_My.md, i have forceField = "Amber";). So i will have to define my own MyForceField.frc file (or i will update Amber.frc file),where i will define "My" atom and whole force field defined in my this rigid body (probably define atom atom "My" and its interactions with others, e.g. interaction between atom "My" and "My", "My" and others, and so). Is it right?
>>>>>>>>
>>>>>>>> If yes, then as first task, i would need to configure this rigid body to be fixed in its initial positions. Can you help me to set it up?
>>>>>>>>
>>>>>>>> Thank you very much
>>>>>>>>
>>>>>>>> Ciao
>>>>>>>>
>>>>>>>> Martin
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> _______________________________________________
>>>>>>>> Openmd-users mailing list
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> Openmd-users at openmd.org
>>>>>>>> http://openmd.org/mailman/listinfo.cgi/openmd-users
>>>>>>> _______________________________________________
>>>>>>> Openmd-users mailing list
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> Openmd-users at openmd.org
>>>>>>> http://openmd.org/mailman/listinfo.cgi/openmd-users
>>>>>> <RB_My.md><RigidBody_01.md>_______________________________________________
>>>>>> Openmd-users mailing list
>>>>>>
>>>>>>
>>>>>> Openmd-users at openmd.org
>>>>>> http://openmd.org/mailman/listinfo.cgi/openmd-users
>>
>>
>>
>> _______________________________________________
>> Openmd-users mailing list
>>
>> Openmd-users at openmd.org
>> http://openmd.org/mailman/listinfo.cgi/openmd-users
>
> _______________________________________________
> Openmd-users mailing list
> Openmd-users at openmd.org
> http://openmd.org/mailman/listinfo.cgi/openmd-users
***********************************************
J. Daniel Gezelter
Associate Professor of Chemistry
Department of Chemistry and Biochemistry
251 Nieuwland Science Hall
University of Notre Dame
Notre Dame, IN 46556-5670
phone: +1 (574) 631-7595
e-mail: gezelter at nd.edu
web: http://www.nd.edu/~gezelter
************************************************
More information about the Openmd-users
mailing list