Thank Dear Axel for your quick respond.
The older and newer versions are lammps-64bit-20160720 and
lammps-64bit-20160902 respectively.
I solved the problem by calculating angular velocities by adding the
following lines to the input file (only the z-component is shown here):
compute COM_Cu Cu com
variable dx atom x-c_COM_Cu[1]
variable dy atom y-c_COM_Cu[2]
variable dz atom z-c_COM_Cu[3]
variable Lz atom (-vx*v_dy+vy*v_dx) # Angluar
momentum in z direction
compute Lz Cu reduce sum v_Lz
variable Iz atom (v_dx^2+v_dy^2) # Moment of
inertia about z-axes
compute Iz Cu reduce sum v_Iz
variable W23 equal c_Lz/c_Iz # angular
velocity of Cu around z-axes
variable wz_NP equal omega(Cu,z) # angular
velocity of Cu around z-axes (by omega function)
compute w_NP Cu omega/chunk CHUNK # angular velocity of Cu
around z-axes (by omega/chunk compute)
The results show complete agreement between the results of the added lines
and both "compute omega/chunk" and "function omega()" which both show
exactly “zero” in the newer version. As you can see the angular velocities
are small but are not "zero":
step
c_w_NP[2][3]
v_wz_NP
v_W23
0
1.46E-06
1.46E-06
1.46E-06
100
2.44E-06
2.44E-06
2.44E-06
200
7.17E-07
7.17E-07
7.15E-07
My original input file is very big so I send a small one to demonstrate the
problem.
Unfortunately, when my simulation box gets higher than a certain value, all
older versions of LAMMPS crash and show this message: "Imp-mpi.exe has
stopped working"
Fortunately, the newer version has not this problem.
Best regards
Sent with