LAMMPS WWW Site - LAMMPS Documentation - LAMMPS Mailing List Archives
Re: [lammps-users] Polymer/CO2 system not time integrating when using fix rigid/nve/small
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [lammps-users] Polymer/CO2 system not time integrating when using fix rigid/nve/small


From: Bill K Wheatle <bkw686@...4851...>
Date: Mon, 14 May 2018 12:30:26 -0500

Hi Axel,

So I took a look at your new version of the script once I recompiled LAMMPS with the new tolerance. It seems that your version certainly got LAMMPS to recognize that the CO2 molecules were indeed linear. However, it seemed that the temperature diverged quite badly. I did manage to combine your edits to input with some additions. (1) I added in a minimization step that removed a good chunk of any overlaps in the system. (2) I assigned velocities right off the bat for the desired final temperature. This seems to run stably for me. I've attached it below for anyone else who may encounter a similar issue in the future.

group           co2 type 1 5
group           poly subtract all co2

print                          .
print ==========================================
print "step1: minimization: 5000 steps CG Minimization"
print ==========================================
print                           .

dump            dumptrj all atom 10 ${molname}_min-cg.trj
dump_modify     dumptrj image yes scale yes
thermo          100
thermo_style    custom time temp pe etotal press lx ly lz density
restart         10 ${molname}1_sd.min ${molname}2_sd.min
min_style       sd
minimize        1.0e-10 1.0e-10 100000 1000000
min_style       cg
restart         10 ${molname}1_cg.min ${molname}2_cg.min
minimize        1.0e-10 1.0e-10 100000 1000000
write_data      data.${molname}_minimize
undump          dumptrj

print =====================================
print "Step2: Heating: NVE dynamics with Langevin thermostat"
print =====================================
print            .

reset_timestep  0
timestep        2.0
fix             fixshake poly shake 0.0001 20 0 b 3 7
fix             fixnve poly nve
fix             fixrigid co2 rigid/nve/small molecule
run             0
unfix           fixrigid
unfix           fixnve

fix             fixrigid co2 rigid/nvt/small molecule temp ${ftemp} ${ftemp} 100.0
fix             fixnvt poly nvt temp ${ftemp} ${ftemp} 100.0
velocity        all create ${ftemp} 12355
run             0
velocity        all scale ${ftemp}
thermo          100                                                                                                                                                                          
thermo_style    custom time temp pe etotal press density
...
run           50000

Thanks for your help.

All the best,

Bill K. Wheatle | Doctoral Candidate
The University of Texas at Austin | McKetta Department of Chemical Engineering | Cockrell School of Engineering
200 E. Dean Keeton St. Austin, TX 78712 | CPE 3.472 | bkw686@...490...851... 


On Sun, May 13, 2018 at 5:14 PM, Axel Kohlmeyer <akohlmey@...24...> wrote:


On Sun, May 13, 2018 at 6:03 PM, Bill K Wheatle <bkw686@...4851...> wrote:
Hello Axel,

Before I use the modified input script, I have a couple of questions. First, what do you mean "precise"?

​you may need to use more digits for the coordinates in the data file​.
 
Is the CO2 angle not 180 degrees in my data file?

​apparently not enough for the code to recognize it.​
 
If not, then would simply minimizing the molecule without any pair interactions turned on help in the future?

​never tried that.​ it might work.
 
Second, what did you change your TOLERANCE setting to in the fix rigid source code?

​form 1.0e-6 to 1.0e-5

axel.

Thanks again for your help.

All the best,

Bill K. Wheatle | Doctoral Candidate
The University of Texas at Austin | McKetta Department of Chemical Engineering | Cockrell School of Engineering
200 E. Dean Keeton St. Austin, TX 78712 | CPE 3.472 | bkw686@...5752...1... 


On Fri, May 11, 2018 at 5:21 PM, Axel Kohlmeyer <akohlmey@...24...> wrote:
bill,

thanks for providing illustrative and complete input examples. that helps a lot in figuring out what might be a problem.

i am seeing two issues:
1) your CO2 geometries are not "precise" enough, so i had to change the TOLERANCE setting in the rigid fix
​ source code and recompile ​
for it to accept those as linear molecules.
2) you have a rather high potential energy initial configuration and there seems to be some conflict with CO2 molecules in combination with the langevin settings, which then makes the time integrator become unstable.

​when i change the middle part of your input to do just one step without any progressions, everything looks like it is straightening out and then a stable MD seems to be following.
see below for the modified input and please give it a try.

axel.​


print                          .
print =====================================
print "Step2: Heating: NVE dynamics with Langevin thermostat"
print =====================================
print            .

reset_timestep  0
timestep        2.0
fix             fixshake poly shake 0.0001 20 0 b 3 7
fix             fixnve poly nve
fix             fixrigid co2 rigid/nve/small molecule
run             0
unfix           fixrigid
fix             fixlangevin poly langevin ${itemp} ${ftemp} 100.0 699483
fix             fixrigid co2 rigid/nvt/small molecule temp ${itemp} ${ftemp} 100.0
thermo          100
thermo_style    custom time temp pe etotal press density



On Thu, May 10, 2018 at 4:09 PM, Bill K Wheatle <bkw686@...4851...> wrote:
Good afternoon,

I am trying to simulate a system of polymer mixed with some CO2 molecules in the 16Mar18 version of LAMMPS with 24 MPI threads. I've attached the input file and the simplified data file containing 1 polymer chain and 20 CO2 molecules. In short, when I apply the rigid/nve/small or rigid/small fixes, the LAMMPS output stops at Timestep: 2.0 without ever showing any thermodynamic data. In contrast, if I simply delete the CO2 from the system and remove the fix, thermodynamic data is produced quite easily. Are there any particular commands that I may be missing in my input file for a mixed system like mine to work correctly?

All the best,

Bill K. Wheatle | Doctoral Candidate
The University of Texas at Austin | McKetta Department of Chemical Engineering | Cockrell School of Engineering
200 E. Dean Keeton St. Austin, TX 78712 | CPE 3.472 | bkw686@...4851... 


------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
lammps-users mailing list
lammps-users@...655....net
https://lists.sourceforge.net/lists/listinfo/lammps-users




--
Dr. Axel Kohlmeyer  akohlmey@...24...  http://goo.gl/1wk0
College of Science & Technology, Temple University, Philadelphia PA, USA
International Centre for Theoretical Physics, Trieste. Italy.




--
Dr. Axel Kohlmeyer  akohlmey@...24...  http://goo.gl/1wk0
College of Science & Technology, Temple University, Philadelphia PA, USA
International Centre for Theoretical Physics, Trieste. Italy.