Please reply to the list, not to me.
Again, you should read the fix deform doc
about the “remap” keyword, which is used
differently for solids vs liquids. I think if
you use remap yes for solids, then it does
not make sense to use fix nvt/sllod. If you
used remap no for solids it could make sense,
but I don’t see why it would make sense to have
a solid “flow” with a deforming box.
Steve
Thank you very much.
When I use "remap none” LAMMPS does not allow to use NVT/SLLOD which is specific for fluid.
So, I use
fix 2 all deform 1 xy variable v_displace v_rate remap none
fix 1 all nvt temp 0.01 0.01 0.1 tchain 3
compute myTemp all temp/deform
fix_modify 1 temp myTemp
As the system does not follow fox deformation, the lower layer is not fixed. It has opposite directional velocity as compared to the top layer. When a small strain amplitude is applied then upper and lower layer will have almost similar velocity with opposite direction. Can it be a reason of some strange values of MSD calculations for low strain amplitude? Though the system does not follow exactly the box deformation, the strain will be calculated as the maximum deformation of the box divided by the y directional box length. Is it right?
Here, I have attached box deformation rate and tow layer velocities for three different strain amplitude (Strain amplitude=0.006, 0.03,0.06 ) to show that top layer and bottom layer moves with closely similar speed as compared to others.
f_0-1.pdf (23.7 KB)
f_0-5.pdf (20.7 KB)
f_1-0.pdf (20.3 KB)
Dear Steve,
I understood one can not expect that a solid will follow the exact box deformation.
From the documentation, for solid one should use fix deform command with remap x. However, when I use remap x and compute temp/deform to subtract the bias i.e,
fix 2 all deform 1 xy variable v_displace v_rate remap x
fix 1 all nvt temp 0.01 0.01 0.1 tchain 3
compute myTemp all temp/deform fix_modify 1 temp myTemp
then it shows the warning
WARNING: Using compute temp/deform with inconsistent fix deform remap option (…/compute_temp_deform.cpp:76)
If you make some comment it would be nice.
From your previous mail, if I understood correctly your suggestion is to use fix deform with remap “none” and fix NVT for solid. Could you please explain if remap is “none", how do the atoms follow box deformation? Because for remap x atoms follow the box deformation due to the position remapping and for remap v, the atoms acquire the box deformation velocity.
One can not use remap “none” and fix NVT/SLLOD because we must need to remap velocity when atoms cross periodic boundary. Is it right?
I am eagerly waiting for your reply.
With best regards,
Pritam
When you use fix deform remap x (for a solid), the positions
of the atoms are remapped to the deformed box shape each
timestep. But even though in a movie it would look like the
atoms are moving with the box, they are not flowing with a velocity.
They have no
streaming velocity that corresponds to the box shape change.
In fact their velocity could be 0.0, if they started with zero
velocity (a cold solid), and would remain so until some kind
of dislocations appeared or plastic deformation took place.
For that kind of model fix nvt/sllod is not appropriate (no thermostat
is even needed), since it will try to force the atoms to have a
streaming velocity, in addition to the position remapping that
is taking place. Hence the warning you are getting.
Steve
Dear Steve,
Thank you very much. Your discussions are quiet helpful.
I am bit confused about the difference between fix deform with "remap none” and “remap v”. In the 2nd case when atoms cross periodic boundary velocity is remapped whereas in the 1st case when atoms cross periodic boundary the velocity of the atoms is not remapped, i.e., it appears on the other side of the box (following Lees-Edwards boundary condition) with the same velocity. However, in both cases, atoms try to follow the box deformation velocity. If I am wrong could you please explain it in short. Because I am not fully understood whether I should use fix deform with remap v or remap none and fix NVT where we study plastic deformation of solid.
If it is possible to send some literatures where this type of system is studied by LAMMPS, that would be nice.
Thanks in advance.
With best regards,
Pritam
For a derforming box, atoms on either side of a periodic boundary can
have different deformation velocities. E.g. for shearing in the x direction,
if an atom crosses the y periodic boundary (e.g. for a fluid) then the
x velocity of the atom should be different at the bottom vs top of the box.
Remap v does that. It adds/subtracts the deformation x velocity to the
atom when it crosses the y boundary. Remap none does not do that (it does
nothing).
I suggest you re-read the section about remap on the fix deform doc page.
It has lots of details and suggested usage tips.
Steve
Dear Steve,
Thanks again. I understood the point.
From the documentation it is clear that fix deform with remap v is appropriate for fluid. So, remap none with fix nvt is reasonable for a solid system.
Another way, I was thinking for solid as follows:
fix 2 all deform 1 xy variable v_displace v_rate remap x
fix 1 all nve
fix 3 all temp/rescale 1 0.01 0.01 0.001 1.0
run 500
In this way, we have affine deformation of the system in every step and MD simulation is performed and temperature is also controlled around 0.01.
I do not see any error message or warning.
Does it make sense to you? It would be nice if you make some comment.
With best regards,
Pritam
I’ve told you everything I can relative to these Qs and how
fix deform and its options work. You’ll have to decide
what makes sense for your model.
Steve
Dear Steve,
Thank you for all discussions. It was extremely helpful. The last one I was asking just because of by this model, I do not find any unexpected behaviour of MSD calculation.
Best regards,
Pritam