Proper use and transition of operational roles

I would like to ask how operational roles should be properly used and most importantly properly transitioned to system analysis. I am designing test system, which will interact with 2 versions of the same system under test (Engineering Model and Flight Model). Accordingly I created 2 hierarchies of operational entities, one for each model reflecting which parts are present in which model. It seems appropriate to me, as subsystems not present in system under test need to be emulated by designed test system. In order not to repeat operational actions and exchanges for these 2 models, I have created roles for different subsystems and assigned them to operational entities in both models.
For OA this approach seems to be working and model validates correctly. After I transition to system analysis, there is only single functional action created for each operational action and they are assigned to multiple system actors, which causes validation error. What would be the correct way to use roles in this case? Are these appropriate at all for my use case? The roles are not covered in example projects like IFE, so it is difficult for me to infer proper usage for them.
Best regards,

Maybe I will rephrase it a little bit: What is the difference in use case of Roles vs. REC/RPL apart from roles being available only in Operational Analysis?

Hi Pawel,
from the Arcadia method point of view (which is my first playground), you are right, roles and REC/RPL answer the same need: defining once a set of model elements that can be reused in various usages, each usage existing in the model separately (so as to be connected, allocated… to its own dedicated environment).
From the Capella tool point of view, roles have some limitations that might lead you to favour REC/RPL instead, but I leave Capella team elaborating on this.
Kind regards.