Hi all
I am trying to simulate a system which contains charged polymers in SPC/E water. I have arranged the atoms initially in random positions. As I run minimization in Lammps (28Jun14) this is what happens. e_coul is suddenly blowing up. But during the first few steps as you can see ecoul has fair values and why does it blow up in middle of the minimization (CG : 1.0e-6 1.0e-8 1000 10000)?
My first data file for this simulation had showed error in first 50steps but after random arrangement this happened. What modification should I do to avoid this?
Step Temp E_pair E_coul PotEng TotEng
0 0 2.3566365e+13 689.98904 2.3566365e+13 2.3566365e+13
50 0 35747.606 814.00225 56125.889 56125.889
100 0 -19308.772 243.87669 -18804.682 -18804.682
150 0 -44931.908 98.042125 -44895.831 -44895.831
200 0 -52907.135 40.986849 -52880.848 -52880.848
250 0 -56527.974 27.993641 -56492.356 -56492.356
300 0 -58280.637 21.7288 -58248.9 -58248.9
350 0 -59690.946 17.497562 -59680.973 -59680.973
400 0 -60536.432 14.060267 -60517.144 -60517.144
450 0 -61021.949 12.515505 -61009.915 -61009.915
500 0 -61616.083 9.2007819 -61607.595 -61607.595
550 0 -61978.977 7.2699535 -61974.464 -61974.464
600 0 -62338.751 3.2325108 -62321.259 -62321.259
650 0 -62685.334 -0.82784055 -62681.032 -62681.032
700 0 -63052.103 -23.50653 -63048.661 -63048.661
739 0 -6.6091591e+16 -6.6091591e+16 -6.6091591e+16 -6.6091591e+16
Loop time of 250.279 on 8 procs for 739 steps with 62519 atoms
Thanks in advance!
Hi all
I am trying to simulate a system which contains charged polymers in SPC/E
water. I have arranged the atoms initially in random positions. As I run
minimization in Lammps (28Jun14) this is what happens. e_coul is suddenly
blowing up. But during the first few steps as you can see ecoul has fair
values and why does it blow up in middle of the minimization (CG : 1.0e-6
1.0e-8 1000 10000)?
because a minimization is in general not sufficient to "unoverlap"
randomized positions.
you should use fix adapt in combination with some soft core potential
(either "soft" or one of the soft-core pair styles in USER-FEP). see
the micelle example for how to set this up.
axel.
Thank you very much for the quick suggestion.
As I was trying to do what you said the water molecules were going through the wall atoms to outside. I used following set of commands.
pair_style soft 16.0
pair_coeff * * 0.0 9.8
pair_coeff 1 3 0.2 9.8
variable prefactor equal ramp(0.0,0.1)
fix 1 flow adapt 1 pair soft a * * v_prefactor
Needless to say, I tried
pair_style soft 16.0
pair_coeff * * 0.0 9.8
variable prefactor equal ramp(0.0,0.1)
fix 1 all adapt 1 pair soft a * * v_prefactor
also. I used 0.1 for prefactor because in my subsequent simulations with LJ potential eps ~ 0.1. And I used NVT for only flow group so that wall atoms wont move. Still same thing is happening.
type 3 is wall and 1 is oxygen in water. So according to the commands shouldn’t the water molecules repel wall and stay inside the channel instead of moving out “through” the wall? Im using real units and wall lattice constant is 5.4.
Thanks in advance.
Thank you very much for the quick suggestion.
As I was trying to do what you said the water molecules were going through
the wall atoms to outside. I used following set of commands.
pair_style soft 16.0
pair_coeff * * 0.0 9.8
pair_coeff 1 3 0.2 9.8
variable prefactor equal ramp(0.0,0.1)
fix 1 flow adapt 1 pair soft a * * v_prefactor
you scale the potential height for *all* atoms. LAMMPS is doing what
you ask for. if you don't want some atoms not to be scaled, you have
to use a different pair style for those (via pair style hybrid or
hybrid/overlay). or just use a wall fix.
Needless to say, I tried
pair_style soft 16.0
pair_coeff * * 0.0 9.8
variable prefactor equal ramp(0.0,0.1)
fix 1 all adapt 1 pair soft a * * v_prefactor
also. I used 0.1 for prefactor because in my subsequent simulations with LJ
potential eps ~ 0.1. And I used NVT for only flow group so that wall atoms
wont move. Still same thing is happening.
it is not about which atoms don't move. it is about how potentials get scaled.
If you are using fix wall during a minimization, this doc page note
is important:
IMPORTANT NOTE: If you want the atom/wall interaction energy to be included in the total potential energy of the system (the quantity being minimized), you MUST enable the fix_modify energy option for this fix.
Also, you should note that constraints, e.g. fix shake for SPC/E are not
enforced during a minimization.
Steve