Dear all,
I am perfoming a compression on a sample that must reach mechanical equilibrium (sum of total forces per particle close to zero). By looking at the fmax variable that can be used in thermo custom I realized that although the system is pretty well equilibrated "visually" and some measurementes (low kinetic energy, number of contacts per atom constant, velocity per particle very small, etc) still the fmax variable is printing large values (it should be close to zero since the sum of force per dimension per particle must be small for a system in mechanical equilibrium). After visualizing the system I realize that the particles with those apparent large forces are those in contact with the walls, so it is like the particles are not storing the wall force when fmax is computing. I am used the following to compute the force per particle:
compute fatom all property/atom fx fy fz
variable ftot atom sqrt(c_fatom[1]*c_fatom[1]+c_fatom[2]*c_fatom[2]+c_fatom[3]*c_fatom[3])
compute ftot all reduce max v_ftot
But it seems that this compute does not take into account the wall forces.
How can I compute the actual force on particles in contact with the walls? BTW, I am using servo walls with fixed force and the granular wall is created as
fix walls all wall/gran model hertz tangential no_history mesh n_meshes 6 meshes servox2 servoy2 servoz2 servox1 servoy1 servoz1 store_force yes store_force_contact yes
The last two keywords were my desperate try to get the forces out somehow. It seems that I can access the force on the walls by using the auto-generated fix, like f_servox1[1] and adding it to the total force, but I cannot find a way for the actual compute shown to take the walls force into account. This also affects the vtk dump, showing non-vanishing force on the particles in contact with walls (and that is how I realized this).
Thanks in advance for any guidance. I hope not to be ignoring something evident.