Sorry, I have no specific advice to address your specific problem. Those region based walls are always confusing me. Thus I usually use those regions in combination with create_atoms to build walls from explicit atoms. I find that less confusing, since you can “see” where the walls are. Also, there are no problems with edges and overlapping faces. The downside is the additional computational cost, but I personally consider that the lesser worry.
Perhaps somebody else on this list with more practical experience in using LAMMPS with fix wall/region can help you.
Axel.
I’m not sure what’s causing your problem, however you may try the inverse geometry by making the region out of the union of a cylinder and two blocks on each end of the cylinder and specifying side in. I’m not sure if this will work any better, but it’s quick enough to implement.
Dear both,
Thank you for your suggestions. I tried to specify the inverse geometry consisting of the union of 2 blocks and a cylinder for use with fix wall/region. Unfortunately, the problem persists because each of the two blocks have a face overlapping with the flat surface of the cylinder. Although I can specify the flat surfaces at the ends of the cylinder as open, it appears impossible to specify a part of the face of a block to be open. The only two scenarios allowed for in the region command are for the face to be open or closed.
Best,
Karthik
Dear all,
One option could be to correct for the missing force using fix addforce, which allows for the application of additional forces upon atoms in a particular region.
Best,
Karthik