A problem about a model which is a infinite plane with a cylinder hole

Dear lammps-users:
It has took me a whole day to build a model which is a infinite plane with a cylinder hole. I want to study the nucleation of liquid covering the plane.The problem i have encountered is “Particle outside surface of region used in fix wall/region (…/fix_wall_region.cpp:240)”.I have tried several methods,such as increasing the distance between the plane and particles, adding value for epsilon of lj93,but failed.So,I am not sure whether my in.test has some faults or not.I have read that “This can cause particles to feel no force (they don’t “see” the wall) when in one location, then move a distance epsilon,and suddenly feel a large force because they now “see” the wall. In the worst-case scenario, this can blow particles out of the simulation box. Thus, as a general rule you should not use the fix wall/region command with union or interesect regions that have convex points or edges” in the manual.So,i may not be able to finish this job,is that right? could anyone give me any suggestions?Could anyone make any labels on my faults,if the faults do exit?
The following is my in.test:
units real
atom_style atomic
boundary p p fs

variable l equal 6.34
variable r equal 4
lattice fcc $l
region box block 0 20 0 20 0 20
create_box 1 box

mass 1 39.948

lattice fcc $l
region hole cylinder z 10 10 $r 1 10 units lattice
region upper block 0 20 0 20 10 20 units lattice
region liquid union 2 upper hole side in
create_atoms 1 random 14311 23145 liquid

lattice fcc $l
variable w equal $r+0.3
region whole cylinder z 10 10 $w 0.7 INF units lattice side out
region wdown block INF INF INF INF INF 9.7 units lattice side in
region wall intersect 2 wdown whole side out

group upper region upper
group hole region hole
group liquid union upper hole

velocity liquid create 150.00 87287

timestep 2.156

pair_style lj/cut 8.5125
pair_coeff * * 0.2381 3.405 8.5125
pair_modify shift yes

neighbor 0.681 bin
neigh_modify every 10 delay 0 check yes

compute holeT liquid temp/partial 1 1 1

thermo_style custom step c_holeT pe ke etotal vol lx ly lz press
thermo 1000

dump 1 all cfg 1000 atom.*.cfg mass type xs ys zs

fix 1 all wall/reflect zhi 126.50 units box

fix 2 liquid wall/region wall lj93 0.4762 3.405 8.5125
fix_modify 2 energy yes
min_style quickmin
minimize 1.0e-5 1.0e-8 1000 10000

fix 3 liquid nvt temp 150.0 150.0 2156
fix_modify 3 temp holeT
run 50000

I doubt anyone is going to understand your model or its

geometry by reading your complex input script. As

with any complex model, you should build it up

one step at a time and test it one step at a time.

E.g. start with a flat plane (no hole). Add your particles,

Can you run it successfully for a small problem (min, dynamics,

nvt, etc). Then add the hole.

No one is better suited to do this than you.