"Must shrink-wrap piston boundary"

LAMMPS LAMMPS Mailing List Mirror
1

Hi
I am trying to simulate shock wave in polymer using the wall/piston feature. I have an equilriated polymer sample, and I define the x and y direction to be periodic, while the z direction (shock propagation direction) is non-periodic.

#fix walls all wall/piston zlo pos 0.0 0.0 0.0 vel 0.0 0.0 0.0 units box which places the piston at the zlo position (left edge of the simulation box, at rest). I then apply an initial particle velocity (Up)

velocity atom_box set 0 0 -0.02 units box

however, I get an error ERROR: Must shrink-wrap piston boundary (fix_wall_piston.cpp:67). While trying to solve the problem, I camoe across two instances, under this fix that differ in the pdf manual and the html doc explaining the command

1 The face that has the wall/piston must be boundary type ‘m’ (shrink-wrapped with minimum a minimum value).
The opposing face can be any boundary type other than periodic.

and

The face that has the wall/piston must be boundary type ‘s’ (shrink-wrapped). The opposing face can be any boundary type other than periodic.

I have tried both options, but cannot get past the error. Not quite sure if I am over looking something, but need help in trying t

of figure out the source of error.



Thanks

J.NJoroge
2

Hi,

If im correct the “m” boundary means the computational box is at a specified value or above that value. Usually After the passing of the Shock wave the material tends to “compress” (reduce its size) . That’s why the shrink-wrapping is used for the pourpouse of readjusting the size of the computational box . i.e Your compresing the sample .

If the box have a minium value then its useless to define a shrink-wrapping boundary (WHAT IS THE POUROUSE OF HAVING a shrink-wrapping minimum boundary?). I would really like to know, because i’m not able to grasp it. (please correct me if im wrong). The only reason to have a shrink-wrapping with a minimum is for the pourpuse of leaving an empty gap between the sample.

You may use the command wall/reflect instead of wall/piston …

A Salute
Oscar G.

3

Hi,
If the box have a minium value then its useless to define a shrink-wrapping
boundary (WHAT IS THE POUROUSE OF HAVING a shrink-wrapping minimum
boundary?). I would really like to know, because i'm not able to grasp it.
(please correct me if im wrong). The only reason to have a shrink-wrapping
with a minimum is for the pourpuse of leaving an empty gap between the
sample.

"m" boundaries are useful for a number of reasons. here are some:
- you need them for use with fix deposit (can't add new atoms unless
  there is an empty space to put them in)
- you don't want lammps to crash when running with a lot of processors,
  if your simulation box is not exactly at the shrinkwrap size (with "s"
  you may lose atoms on the first re-neighboring)
- you want to steer lammps to go for a specific domain distribution in
  domain decomposition (since you know roughtly the final shape of
   your box)
- you may want to keep the box at a minimum size for visualization
  purposes.

axel.

4

Axel,

Thanks for answering my Doubt and your valuable feedback =) . Althought I still dont know what would be the benefit of using a Shitnk-wrapping minimum “sm” boundary
for the pourpouse of simulating a shock wave phenomena …

A salute
Oscar G

5

Is there a reason why you recommend wall/reflect as opposed to wall/piston?