Error, shake vs fix smd (PMF)

I want calculate PMF for adsorption of a molecule on surface of graphene in a molecule-graphend-water collection.

I use fix shake for only water molecules. Graphene is the fixed group and Molecule is the free group in calculating the PMF using fix smd. When I use fix smd with couple and cfor mode, I get no error (although the value is 0!!). But when use couple and cvel it stops with an error after alot of warnings (see figure):

Warning: Improper problem: %d %ld %d %d %d %d

Error on proc 0: Shake determinant = 0.0 <…/fix_shake.cpp:1969)

I’m using these commands in LAMMPS (30 Sep 2014-ICMS):

fix sh all shake 0.000001 1000 0 t 2 3 b 2 a 2 # on water molecules
fix Eq1 all npt temp 300 300 100 iso 1 1 1000

fix PMF Graphene smd cvel 20 0.0001 couple Mol1 auto auto auto 0
fix PMF all ave/time 1 500 500 f_PMF[1] f_PMF[2] f_PMF[3] f_PMF[4] f_PMF[5] f_PMF[6] f_PMF[7] file tmp.pmf

What is the problem of using cvel in my simulation?
Thank you.

error.jpg

I want calculate PMF for adsorption of a molecule on surface of graphene in
a molecule-graphend-water collection.

I use fix shake for only water molecules. Graphene is the fixed group and
Molecule is the free group in calculating the PMF using fix smd. When I use
fix smd with couple and cfor mode, I get no error (although the value is
0!!). But when use couple and cvel it stops with an error after alot of
warnings (see figure):

Warning: Improper problem: %d %ld %d %d %d %d

Error on proc 0: Shake determinant = 0.0 <../fix_shake.cpp:1969)

I'm using these commands in LAMMPS (30 Sep 2014-ICMS):

fix sh all shake 0.000001 1000 0 t 2 3 b 2 a 2 # on water
molecules
fix Eq1 all npt temp 300 300 100 iso 1 1 1000

fix PMF Graphene smd cvel 20 0.0001 couple Mol1 auto auto auto
0
fix PMF all ave/time 1 500 500 f_PMF[1] f_PMF[2] f_PMF[3]
f_PMF[4] f_PMF[5] f_PMF[6] f_PMF[7] file tmp.pmf

What is the problem of using cvel in my simulation?

there is not enough information here to say anything but, there is
something wrong in your simulation setup.
have you done the obvious step and visualized your simulation and
checked what is happening? have you checked which molecule the
"problematic" impropers belong to?

axel.

Thank you for your reply.

Upon checking, I found the problematic impropers are relate to molecule and graphene both.

I visualized it, and couldn’t find a problem.

When the simulation run without fix smd, no error or warning is shown. It work good.

When the simulation run with fix smd, but without shake command, it doesn’t stop with error, but the warnings appear again. But, there is no progress in simulation. No progress in steps.

I attached the files for you.

I wonder how the fix shake commands on only water molecule can interfere with impropers of graphene and molecule and/or with pmf calculation of molecule and graphene. It seems so complicated.

Molecule.data (462 KB)

Molecule.in (2.08 KB)