We are also wrestling with this question and have posted my query to another thread in the PID forum.
See: Help! PIDs for facilities and instruments
Basically we have come to a number of conclusions.
First is that at this time there doesn’t appear to be a consensus on where to register research infrastructure components such as laboratories, research stations, instruments and other physical facilities supporting research.
Second, we have conceptually (and perhaps arbitrarily) divided research infrastructure into physical objects like instruments or vessels, sattelites, telescopes, etc. and organizational components such as research stations and laboratories, which may themselves contain multiple instruments or other more physical infrastructure components.
For the instruments, we have decided to use DataCite (of which we are members) to create PIDs. DataCite includes a ‘Physical Object’ as type and that will suffice. Our initial use will include a high performance computing cluster and other specialized equipment.
For biological research stations, we are waiting for a best practice to emerge. The Research Organization Registry does not yet seem capable of managing what could be a high volume of these organization research infrastructure components but we are hopeful that sometime soon there will be a policy on this. For example, in the ROR there is the Smithsonian Institution, the Smithsonian Tropical Research Institute (in Panama). We would like to create a PID for the Barro Colorado Island research station which is part of the Tropical Research Institute in Panama. But it is not clear that ROR is ready for this.
So at this point we may just use a local identifier or the Q number in WikiData or the VIAF identifier or something similar for our immediate purposes. When a more universally-accepted solution emerges, we are eager to adopt.
I hope this helps. Please be sure to let us know what flavor of PID you end up using.
Alvin Hutchinson