[lammps-users] A question about "displace_box"

Hi, Steve:
I set up a simple orthogonal box whose three parameters are 1.81nm*1.81nm*5.05nm, the atoms are all Cu. I want to use the "displace_box" to give the box an initial strain in z direction. My in.* file is as below:

# 3d Cu100
#Initialization
units metal
atom_style full
#Atom definition
read_data data.Cu100
#Force-field
pair_style eam
pair_coeff * * cuu3.eam
#Settings
group 1 id <= 100
group 2 id >= 1301
group 3 union 1 2
group 4 subtract all 3
variable l_z equal lz
variable lz0 equal \{l\_z\} variable lz3 equal v\_lz0/100 minimize 0 0 100000 100000 reset\_timestep 0 fix 1 all npt 0\.001 0\.001 100 aniso 0 0 0 0 NULL NULL 1000 drag 1 run 5000 reset\_timestep 0 compute 1 all stress/atom compute 2 all reduce sum c\_1\[3\] variable stress equal c\_2/\(vol\*10^4\) variable strain equal \(lz\-v\_lz0\)/v\_lz0 variable ratio equal lx/v\_lx0 fix 5 3 setforce 0 0 0 displace\_box all z delta 0 {lz3} units box
fix 3 all ave/time 1 200 1000 v_stress
thermo_style custom step v_strain f_3
thermo_modify flush yes
thermo 1000
run 71000

I am sure that the data.Cu100 and cuu3.eam are all OK, Because I have tried to change "displace_box" with "fix deform" command, and the data the programme gives accords with the experiment well. But what make me confused is that if I used the "displace_box" command, the mean stress of all the atoms in z direction is unstable, no matter how many times I run after the initial strain. Here is some part of my log file.
Step strain 3
       0 0.01 0
    1000 0.01 0.53821053
    2000 0.01 0.83189227
    3000 0.01 0.35370564
    4000 0.01 1.0049092
    5000 0.01 0.19629893
    6000 0.01 1.1434605
    7000 0.01 0.079951733
    8000 0.01 1.2353194
    9000 0.01 0.0148101
   10000 0.01 1.2725684
   11000 0.01 0.0063304946
   12000 0.01 1.2522789
   13000 0.01 0.054829905
   14000 0.01 1.1767381
   15000 0.01 0.15549704
   16000 0.01 1.0532004
   17000 0.01 0.29886284
   18000 0.01 0.89319448
   19000 0.01 0.47168094
   20000 0.01 0.71145001
   21000 0.01 0.65814357
   22000 0.01 0.52454075
   23000 0.01 0.84132553
   24000 0.01 0.34936956
   25000 0.01 1.004723
   26000 0.01 0.2016375
   27000 0.01 1.1337449
   28000 0.01 0.094437803
   29000 0.01 1.2170181
   30000 0.01 0.037102365
   31000 0.01 1.2473888
   32000 0.01 0.034398824
   33000 0.01 1.2225344
   34000 0.01 0.086140292
   35000 0.01 1.1451371
   36000 0.01 0.18724051
   37000 0.01 1.0226009
   38000 0.01 0.32820742
   39000 0.01 0.86634678
   40000 0.01 0.49602504
   41000 0.01 0.6907501
   42000 0.01 0.67534836
   43000 0.01 0.51181711
   44000 0.01 0.849905
   45000 0.01 0.34572268
   46000 0.01 1.0039743
   47000 0.01 0.20734638
   48000 0.01 1.1238069
   49000 0.01 0.10894142
   50000 0.01 1.1988536
   51000 0.01 0.059056433
   52000 0.01 1.2226925
   53000 0.01 0.061801298
   54000 0.01 1.1935768
   55000 0.01 0.116514
   56000 0.01 1.1145565
   57000 0.01 0.21785723
   58000 0.01 0.99316456
   59000 0.01 0.35633292
   60000 0.01 0.84070142
   61000 0.01 0.51916396
   62000 0.01 0.67118453
   63000 0.01 0.69146726
   64000 0.01 0.50005855
   65000 0.01 0.85761293
   66000 0.01 0.34278533
   67000 0.01 1.0026446
   68000 0.01 0.21344566
   69000 0.01 1.113629
   70000 0.01 0.12348099
   71000 0.01 1.1808106

What is the difference between "displace_box" and "fix deform" on change the shape of the box, except that the later is used for dynamics run?
Thank you~
Regards
Xu

I don't know what you are asking about your simulation results.
But the difference between the 2 commands is:

displace_box will change the box size/shape instantly to the
new value

fix deform will do it gradually over the course of a dynamics run

Steve