Dear Users,
I am simulating a system of 2D polymers in LAMMPS. But I am not able to fix the problem of FENE bond too long and the simulation results in segmentation fault (core dumped). I have tried with varying the parameters in bond_coeff commands and the cut off values etc. But still the problem remains as such. I have a box dimension of 50x50x0.5. I have set my Z dimensions narrow as described in LAMMPS documentation. So how can I fix this issue…
Below is my input script:
…
dimension 2
units lj
boundary p p p
atom_style hybrid bond sphere
neighbor 0.3 bin
neigh_modify every 1 delay 10
read_data system_poly2D.data
bond_style fene
bond_coeff 1 40.0 3.0 0 0
bond_coeff 2 80.0 1.0 0 0
special_bonds fene
pair_style lj/cut 2.5
pair_modify shift yes
pair_coeff 1 1 1.0 1.0 1.22
pair_coeff 2 2 1.0 1.0 1.22
pair_coeff 1 2 1.0 1.0 1.22
pair_coeff 1 3 1.0 1.0 1.22
pair_coeff 2 3 1.5 1.0 1.22
pair_coeff 3 3 1.0 1.0 1.22
minimize 1e-5 1e-5 10000 10000
velocity all create 1.0 234985 rot yes dist gaussian
fix 1 all nve
fix 2 all langevin 1.0 1.0 10.0 954297
timestep 0.001
thermo 10
run 10000
…
Please help me to resolve this issue…
Thanks in advance
-Rahul
I suggest you look at the script bench/in.chain
and its data file. That will show you typical
values for FENE bond coeffs. Also you
should not use atom style hybrid bond sphere.
Just atom_style bond. With point particles.
Steve
But the problem is that, I am using bond style for polymers in my simulation system and the sphere style stands for the spherical colloidal particles present within the same simulation system. So I need to use a hybrid style here. I have used sphere style to make it a spherical particle (then only I can account for rotational motion) otherwise it will also consider as a point particle in LAMMPS. So is the issue is with hybrid style or the 2D setup…
Thank You
Rahul
I suggest you look at the script bench/in.chain
and its data file. That will show you typical
values for FENE bond coeffs. Also you
should not use atom style hybrid bond sphere.
Just atom_style bond. With point particles.
Steve
There is nothing in your input script that
will use finite-size spheres or rotating
particles. You are using pair lj/cut and
bond fene. I suggest you first build
a model for polymers only (similar to bench/in.chain)
and verify you can run correct dynamics. Then
add some larger colloid particle (just with large sigma?),
not bonded to the polymers, and verify correct
dynamics. Then (if this is your model), add bonds
between the colloid and ends of polymer chains,
and test dynamics. You will certainly need different
FENE params between 2 polymer beads
versus a large sigma colloid and a bead. You can’tjust make them up, You would need to study the FENE
formula (and original paper) and figure out correct
params to choose. Again a atom_style sphere
is not needed for any of this. If you need it for
some other reason, then do each of the 3 steps with
a hybrid atom style and verify they work, one at a time.
Steve
Thank you Steve for your explanation… Here I am simulating a system in which the polymers and colloidal particles are not bonded to each other. The bonds is between the polymer beads and between two colloidal particles only. The polymer and colloidal particles are not bonded to each other. As you mentioned I will check the FENE reference paper. My query is that how can I calculate the FENE parameters for a system without fixing either the K value or ‘r’ value. I have simulated something similar to this in 3D. There is no problem with the FENE bond when I used K=40 and r=3.0. So is it the issue with dimension…
Thank You
-Rahul
The output for a FENE bond too long tells which 2 particles
the bond is between. Are they polymer monomers, or 2 colloid
particles? If the latter, I suggest you run a system with
only a few bonded colloids and see if it is stable. Also,
you should look at both the fene and fene/expand
doc pages and understand and plot the formulas for
the bond potential. If you are bonding sigma != 1.0 particles
with FENE you have to adjust the params accordingly.
Reading the original FENE paper (Kremer, Grest) would
also help you.
Steve
I think the error is due to the fact that he is not using the fix enforce2d in his script. May be that is why the error is appearing repeatedly.
-Ligesh