Problem about Error: Invalid thermo keyword in variable formula

LAMMPS LAMMPS Mailing List Mirror
1

Dear all,

I’m trying to use compute heat/flux command to calculate thermal conductivity of Cu-Ar nanofluids while facing an error which goes :

ERROR: Invalid thermo keyword in variable formula (…/variable.cpp:2008)

I’ve checked my program and found no error. Besides, i notice that while the fix ave/correlate command was annotated, the error disappeared.(although the error messege says error relates to the variable formula) I’m confused about this problem and i appreciate any advice.

Part of my program is as follows: (I’m using LAMMPS (8 Jun 2016-ICMS))

thermpal conductivity calculation

compute ArKe Ar ke/atom

per-atom vector

compute ArPe Ar pe/atom

per-atom vector

compute ArS Ar stress/atom NULL virial

per-atom vector

compute ArFlux Ar heat/flux ArKe ArPe ArS

6 component vector:[Jx Jy Jz Jx_v Jy_v Jz_v]

compute ArVelo Ar property/atom vx vy vz

3 per-atom vectors

variable ArEnthalpy equal ave(c_ArKe)+ave(c_ArPe)

scalar

variable ArfluxX equal c_ArFlux[1]-(v_ArEnthalpy*sum(c_ArVelo[1]))

scalar, partial enthalpy of each atom

variable ArfluxY equal c_ArFlux[2]-(v_ArEnthalpysum(c_ArVelo[2]))
variable ArfluxZ equal c_ArFlux[3]-(v_ArEnthalpysum(c_ArVelo[3]))

compute CuKe Cu ke/atom
compute CuPe Cu pe/atom
compute CuS Cu stress/atom NULL virial
compute CuFlux Cu heat/flux CuKe CuPe CuS
compute CuVelo Cu property/atom vx vy vz

variable CuEnthalpy equal ave(c_CuKe)+ave(c_CuPe)
variable CufluxX equal c_CuFlux[1]-(v_CuEnthalpysum(c_CuVelo[1]))
variable CufluxY equal c_CuFlux[2]-(v_CuEnthalpysum(c_CuVelo[2]))
variable CufluxZ equal c_CuFlux[3]-(v_CuEnthalpy*sum(c_CuVelo[3]))

variable fluxX equal ArfluxX+CufluxX

scalar

variable fluxY equal ArfluxY+CufluxY
variable fluxZ equal ArfluxZ+CufluxZ

fix JJ all ave/correlate $s $p $d &
v_fluxX v_fluxY v_fluxZ type auto &
file profile.heatflux ave running

error occurs here

#variable scale equal s*dt*{convert}/vol/${kB}/$t/t #variable k11 equal trap(f_JJ[3])*{scale}
#variable k22 equal trap(f_JJ[4])*{scale} #variable k33 equal trap(f_JJ[5])*{scale}
#variable kappa equal (v_k11+v_k22+v_k33)/3.0

thermo $d
thermo_style custom step temp density

v_fluxX v_fluxY v_fluxZ v_k11 v_k22 v_k33 v_kappa

dump 4 all xyz 5000 calculation.xyz

run ${step_2}

Best regards,
George

2

Dear all,

I'm trying to use compute heat/flux command to calculate thermal
conductivity of Cu-Ar nanofluids while facing an error which goes :

ERROR: Invalid thermo keyword in variable formula (../variable.cpp:2008)

I've checked my program and found no error. Besides, i notice that while
the fix ave/correlate command was annotated, the error
disappeared.(although the error messege says error relates to the variable
formula) I'm confused about this problem and i appreciate any advice.

​it is easy to be confused by such a complicated construct. it is a bad
idea to write such a mess and then try to figure out what is wrong
afterwards. why don't you just remove everything and put​ it back piece by
piece in stages? then you can easily identify the issue.

according to the error message, the problem is likely to be one of the
variable commands.
my money is on using the ave() functions incorrectly. you apply them to
per-atom computes, but they are meant to be used on global vectors or
arrays. to get the kind of average, you are looking for you would have to
do a reduction first with compute reduce.

axel.

3

Lines like this:
variable fluxX equal ArfluxX+CufluxX

are not correct syntax.

Needs to be:
variable fluxX equal v_ArfluxX+v_CufluxX

LAMMPS is trying to interpret ArfluxX as a thermo

keyword …

Steve