Taking the pressure tensor as stress I am getting quite reasonable stress-strain curve.
I just read your email more carefully. (Kind of embarrassed. Did not
mean to pontificate like that.)
Using the pressure tensor sounds fine to me. (But again, please check
with someone who knows more than me.)
Good luck.
Andrew
Good Morning,
Using Fix NVE/limit I can supersede that error "Pair table inner cut off distance" and I am using fix deform as follows:
fix 1 all nve/limit 0.05
fix 3 all langevin 298 298 1 290
fix 2 all deform 1 x erate 1e-5 units box remap xTaking the pressure tensor as stress I am getting quite reasonable stress-strain curve. Using Fix NVE/limit, can I accept this stress-strain values?
the fact that you are asking this question, is a hint that you are not
100% certain. but you *should* be. whether somebody else says "yes" or
"no" is not really going to solve the problem, that it is *you* that
has to justify the result later.
so how do you get this confidence? you have to "earn" it. i.e. look
for examples, tutorials, publications etc. you are not the first
person doing this kind of study, so you can see how others address it
and you need to try reproducing what others did (even if it is on a
different system). look for simple cases and then work your way up. if
you have technical problems, you are welcome to ask here, but don't
expect too many positive reactions, when you ask for approval of
inputs or results.
that being said, here a few suggestions to the specific details of your problem.
nve/limit (like temp/rescale) should *never* be used in a production
calculation. however, that doesn't mean that you should not use it.
you just need to split your simulation into two parts. one part for
equilibration, one part for production. in the first part, you can use
nve/limit to make sure your system doesn't misbehave too much. but in
the production part it has the potential to completely invalidate your
results, since it manipulates the system in a completely unphysical
way. these kind of tools exist, since the way how initial geometries
are constructed is often as unphysical and you start from an initial
geometry that is too far off the equilibrium.
however, if you run into such a situation (as you are right now),
there are also other, cleaner ways to reduce the problem. you can
start by running a minimization for example "to take the edge off". or
check, if there are ways to construct the initial geometry with less
potential energy, e.g. reduce overlap, find a better match of
parameters etc. there are many ways. as i wrote before, you have to
explore, be alert and learn. discuss with your adviser, discuss with
colleagues (especially senior and experienced ones) all of that is
helping you to earn the confidence, so you won't need to ask questions
like this.
deal?
axel.
Thanks Dr. Andrew for your valuable time and suggestions. I appreciate it.
Dr. Axel, Thank you always. Your suggestion will help me to be self dependent. I know that nve/limit is used for equilibration not for finding final properties. But I was interested to know how strict the professionals are about this issue. Is it that it can never be accepted because there is huge changes? or the quantitative value of the results would be wrong but the behavior can be accepted considering some error? However thanks. I got my answer.
Thanks
Best Regards
Md Salah Uddin