Hello LAMMPS users,
I am trying to perform energy minimization of a system consist of three materials(silica, DGEBA, and water) but it gives following error:
- if I use “region 1 block -0.069857 25.350143 0.219866 27.694866 0.869611 53.704087 units box”
ERROR: Computed temperature for fix temp/berendsen cannot be 0.0 (…/fix_temp_berendsen.cpp:147)
Last command: run 500
- if I use “region 1 block 0 25.42 0 27.47 0 52.835 units box”
ERROR on proc 0: Bond atoms 1005 1007 missing on proc 0 at step 593 (…/ntopo_bond_all.cpp:63)
application called MPI_Abort(MPI_COMM_WORLD, 1) - process 0
I have attached LAMMPS data file and input file for your reference.
Please help me to find out where I’m going wrong.
Is there any way to resolve these error?
Thanks & Regards,
Manish Solanki
new.data (366 KB)
Test.in (1.67 KB)
I find it highly unbelievable that changing a region somehow made the thermostat error go away.
It sounds like bad dynamics to me, but right now I cannot check your script. Search the mailing list for “lost atoms” and you’ll find many questions and answers related to this.
If your dynamics do turn out to be OK, increasing your neighbor skin distance with the “neigbor” command might help.
Possibly
neigh_modify every 1 delay 0 check yes
as well. I’m not sure why this isn’t default values.
Anders
Possibly
*neigh_modify every 1 delay 0 check yes*
as well. I'm not sure why this isn't default values.
it *is* used for minimization. you need to ask the LOTL about why it is
not the default for everything.
axel.
Hello LAMMPS users,
I am trying to perform energy minimization of a system consist of three
materials(silica, DGEBA, and water) but it gives following error:
1) if I use "region 1 block -0.069857 25.350143 0.219866
27.694866 0.869611 53.704087 units box"
*ERROR: Computed temperature for fix temp/berendsen cannot be 0.0
(../fix_temp_berendsen.cpp:147)*
*Last command: run 500*
this is due to the berendsen algorithm. if your system has no temperature,
i.e. no kinetic energy assigned, it cannot work.
your region command creates a group hold_substrate, that contains *all*
atoms. now, since you set for all of these atoms, the velocity to zero and
also zero the forces, you cannot rescale them to a finite temperature.
PEBCAC!
2) if I use "region 1 block 0 25.42 0 27.47 0 52.835 units box"
*ERROR on proc 0: Bond atoms 1005 1007 missing on proc 0 at step 593
(../ntopo_bond_all.cpp:63)*
*application called MPI_Abort(MPI_COMM_WORLD, 1) - process 0*
the situation is similar to above, most atoms are make immobile, but a few
atoms are outside the large region. now you simulation becomes unstable.
this usually means, that you are using a larger timestep than what results
in a stable time integration for your system. you have multiple hydrogen
atoms, but use a time step of 1.0 on a rather unbalanced setup without
constraints on bonds involving hydrogens. that is not likely to result in
stable time integration.
again: PEBCAC!
I have attached LAMMPS data file and input file for your reference.
Please help me to find out where I'm going wrong.
bad model, bad parameters -> crashing simulation.
axel.
Dr. Axel Kohlmeyer,
Thank you for the clarification!
Also, can you please suggest me a tool to create the data file with dihedral info?
I tried to create the data file with CVFF force field parameter by using msi2lmp tool with .car and mdf (generated by Material Studio). but it says Unable to find bond data for cp o. that means there is no bond information for atom cp and O. So how I can add the missing info to force field as the research paper which I’m referring has mentioned these parameters.
I tried to edit cvff.frc file but it resulted in> msi2lmp stop working.
Thanks & Regards,
Manish
Dr. Axel Kohlmeyer,
Thank you for the clarification!
Also, can you please suggest me a tool to create the data file with
dihedral info?
http://lammps.sandia.gov/prepost.html
I tried to create the data file with CVFF force field parameter by using
msi2lmp tool with .car and mdf (generated by Material Studio). but it says *Unable
to find bond data for cp o. *that means there is no bond information for
atom cp and O. So how I can add the missing info to force field as the
research paper which I'm referring has mentioned these parameters.
I tried to edit cvff.frc file but it resulted in> msi2lmp stop working.
then you didn't edit the file correctly: garbage in -> garbage out.
an alternate approach is to tell msi2lmp to ignore such errors (it will
then insert a value of 0.0) and then edit the resulting file in a text
editor and fill in the suitable parameters.
axel.