Hi, I am Yu-Chuan Cheng from NTHU,
I have a problem about twisting balloon.
I have a long cylinder balloon membrane, and I want to twist it.
How can I incorporate air conservation in simulation ?
the following link is the video of twisting balloon,
https://www.youtube.com/watch?v=nG3pErZ2rEg
Best regards
thanks for your help
the lammps code is following:
clear
units lj
boundary p p p # p = periodic
atom_style angle # only bonds and angle
neighbor 0.3 bin # 0.3 = extra distance beyond force cutoff (distance units)
neigh_modify delay 5
variable step equal step
variable k_bond equal 2000000.0
variable k_angle equal 50000.0
variable randomseed equal 239710 # 6位數字
variable epsilon equal 1.0
variable sigma1 equal 1.0
variable sigma2 equal 0.01
variable sigma3 equal 1.0
variable cutoff1 equal ${sigma1} # lj potential cutoff
variable lowerb equal -100.0
variable upperb equal 100.0
variable dumpstep equal 500 # per dumpstep output a dump
variable thermostep equal 500 # per thermostep output a thermo
variable totalstep1 equal 150000
variable timestep equal 0.00001
variable PEStep equal 23000
variable force equal 5.0
variable VELOCITY equal 1.0
variable R0 equal 11
read_data R10L220w.cylinder
region 1 block {upperb} EDGE EDGE EDGE EDGE EDGE
region 2 block EDGE {lowerb} EDGE EDGE EDGE EDGE
region 3 cylinder x 0 0 v_R0 {lowerb}-10 {lowerb}+10 open 2
region 4 cylinder x 0 0 v_R0 {upperb}-10 {upperb}+10 open 1
group 1 region 1
group 2 region 2
group 3 id <= 16758
group 4 id > 16758
pair_style lj/cut {cutoff1} # lj potantial / 1.122 = result of differential
pair_coeff 1 1 {epsilon} {sigma1} {cutoff1} # atom type 2 term [epsilon] [sigma] [cutoff]
pair_coeff 1 2 {epsilon} {sigma2} {cutoff1} # atom type 2 term [epsilon] [sigma] [cutoff]
pair_coeff 2 2 {epsilon} {sigma3} {cutoff1} # atom type 2 term [epsilon] [sigma] [cutoff]
bond_style harmonic # potential = E = K(r-r0)^2
bond_coeff 1 {k_bond} 1.0 # [bond type] [K] [r0]
#angle_style harmonic # potential = E = K(theta-theta0)^2
#angle_coeff 1 {k_angle} 180 # [angle type] [K] [theta0(degree)]
angle_style area_volume
angle_coeff 1 1 1 1 1 1 # type ka a0 kv v0 kl
fix a 1 nvt temp 0.0001 0.0001 1.0
fix b 2 nvt temp 100 100 1.0
#fix w all wall/region 3 lj93 {epsilon} {sigma1} {cutoff1} # [regionID] [epsilon] [sigma] [cutoff]
#fix w all wall/region 4 lj93 {epsilon} {sigma1} {cutoff1} # [regionID] [epsilon] [sigma] [cutoff]
fix left 1 setforce NULL 0.0 0.0
fix right 2 setforce NULL 0.0 0.0
#variable Vyu atom ${VELOCITY}*z/(y^2+z^2)^0.5
#variable Vzu atom -${VELOCITY}*y/(y^2+z^2)^0.5
#variable Vyl atom -${VELOCITY}*z/(y^2+z^2)^0.5
#variable Vzl atom ${VELOCITY}*y/(y^2+z^2)^0.5
#velocity 1 set NULL v_Vyu v_Vzu
#velocity 2 set NULL v_Vyl v_Vzl
compute a 1 property/atom y
compute b 1 property/atom z
compute c 2 property/atom y
compute d 2 property/atom z
variable ycoordu atom -{force}*c_a
variable zcoordu atom {force}*c_b
variable ycoordl atom {force}*c_c
variable zcoordl atom -{force}*c_d
fix aleft 1 addforce 0.0 v_zcoordu v_ycoordu
fix aright 2 addforce 0.0 v_zcoordl v_ycoordl
compute 4 all pe/atom angle # per-atom energy is calculated by the various angle
compute 3 all pe/atom bond # per-atom energy is calculated by the various bond
#dump 1 all custom {totalstep1} pe%.R41L82_Kb{k_bond}_Ka${k_angle}_s${totalstep1} id c_3 c_4
dump 2 all custom ${dumpstep} atomlocal.* id x y z
timestep {timestep}
thermo {thermostep}
thermo_style custom step temp eangle ebond
restart ${totalstep1} twistcylinder.example
run ${totalstep1}
鄭又銓 <s103022207@…24…> 於 2019年11月19日 週二 下午8:07寫道：
I know people have modeled flexible membranes with LAMMPS,
e.g. for closed volumes like a biological cell. These are coarse-grained
models where the membrane is simply beads bonded together, like
a 2d hexagonal lattice of harmonic bonds. If you make the bead separation small enough then
nothing inside will escape. However you don’t need anything inside
to “inflate” the membrane. The fact that it wants to be a 2d plane
is enough to make it balloon like.
Whether you can twist it to create
balloon animal shapes in an MD simulation is a separate Q.
Just apply a spatial/time dependent fold-into-an-elephant external force field
via fix addforce I guess.
Steve