ERROR: Attempting to rescale a 0.0 temperature (../velocity.cpp:715)

LAMMPS LAMMPS Mailing List Mirror
1

Dear LAMMPS users,

I am simulating the granular system by LAMMPS. I have no problem with the LAMMPS (8 Aug 2014), everything is working well; but after I upgrade LAMMPS to the Stable version (10 Feb 2015) , I meet the problem :

ERROR: Attempting to rescale a 0.0 temperature (../velocity.cpp:715)

using the absolute same input script. The command "velocity chain create $T 4928459 mom yes rot yes dist gaussian" kills the program, if I change it to "velocity chain set 0 0 0 units box", it works, but it is not what I want.

I also use LAMMPS to simulate alkane, the atom_style is full, the similar command "velocity all create $T 4928459 mom yes rot yes dist gaussian" initiates the velocities, both versions work.

Following is part of my input script for granular system:

/////////////////////////////////////////////////////////////////
newton off

processors * * 1

variable T equal 1.0 # Temperature

variable A equal 2.0 # Amplitude

variable W equal 0.7 # Period

variable K equal 5.0

variable l equal 3.1

variable DT equal 0.01

timestep ${DT}

thermo 1000

units lj

atom_style sphere

atom_modify map array sort 10000 2.0

dimension 3

boundary p s s

pair_style gran/hooke/history 20.0 NULL 50.0 NULL 0.5 1

read_data system.data

pair_coeff * *

#read_restart restart_before_steady_state.1000000.friction

group chain type 1 2 3

atom_modify first chain

group bot type 4

group bead1 id 1

group bead2 id 2

group bead3 id 3

group bead4 id 4

group bead5 id 5

neighbor 0.3 bin

#neighbor 2.0 nsq

neigh_modify every 5 delay 0 check yes exclude group bot bot

comm_style tiled

comm_modify vel yes

velocity chain create $T 4928459 mom yes rot yes dist gaussian

////////////////////////////////////////////////////////////////////////
Thank you very much!

Best wishes,

Wei

2

Dear LAMMPS users,

I am simulating the granular system by LAMMPS. I have no problem with
the LAMMPS (8 Aug 2014), everything is working well; but after I
upgrade LAMMPS to the Stable version (10 Feb 2015) , I meet the problem :

ERROR: Attempting to rescale a 0.0 temperature (../velocity.cpp:715)

using the absolute same input script. The command "velocity chain
create $T 4928459 mom yes rot yes dist gaussian" kills the program, if I
change it to "velocity chain set 0 0 0 units box", it works, but it
is not what I want.

I also use LAMMPS to simulate alkane, the atom_style is full, the
similar command "velocity all create $T 4928459 mom yes rot yes dist
gaussian" initiates the velocities, both versions work.

due to new features in LAMMPS, initialization has become quite complex
and it appears that in your specific case of granular media
simulation, some setup is missing. this can often be worked around by
inserting a "run 0" statement, which enforces executing a full setup.

axel.

3

There was a bug when read_data shrink-wrapped the box

around the atoms. It didn’t reset the atom map, which

caused no velocities to be assigned in your script. Hence
a temperature of 0.0. It will be fixed in the next patch.

Steve

4

Thank you very much! It works.

Best wishes,

Wei