Why all the atoms are not visible in pps boundary condition?
Because they have moved away or were lost.
ok, so how we can ensure the confinement of atoms within the simulation box? Based on study done in previous research, researchers have used PPS boundary conditions for the similar condition
You will lose atoms them because a) you have chosen to ignore lost atoms and b) there are high potential energy conditions at the beginning or during the simulation. This will be evident from your input and log. They have to contain meaningful and correct choices and show that the simulation has proceeded normally, but - quite likely - that is not the case here.
MD simulations strictly follow the GI-GO principle (i.e. garbage in - garbage out) and it is impossible to say from just looking at visualizations if this applies here.
It appears that the simulation box is initially adjusted based on the atoms defined immediately after setting up the simulation box. I defined all the regions initially, so it worked. However, I am facing challenges in adding atoms later in the process. I am attempting to simulate the melting of a powder bed layer, where atoms are added and melted repeatedly in layers. How can I effectively add atoms to the simulation after the initial setup to simulate the melting of additional layers?
That is the documented behavior of shrinkwrap boundary conditions.
Apparently, shrinkwrap boundaries are not the right choice for your system and workflow.