I would say that it is never wrong to check any viable scenario, but in terms of sizing the only accumulations into the suction drum would be the fresh feed and the power input from the compressor.  The amount of fresh feed coming into the system during a relief case should be known to you. Since the recycle flow to the drum ultimately comes from the drum, there is no additional material accumulation due to the recycle, however, the power input of the compressor could be significant- i.e. in the extreme case of total spillback without an aftercooler, the entire power input will be manifested as a temperature and pressure rise. Even if there were no fresh material entering, the energy input would have to leave the system through the PSV.

These are just my thoughts. Best wishes, sshep

If the recycle valve wide open has greater capacity than the compressor you could get additional backflow from the high pressure system.  That is something that will need to be checked as well as what other safeguards you may have to address backflow.  Even if it turns out that it is not a sizing basis for the relief valve, you might as well include that in you relief system documentation that you looked at it and found it to be no concern.

Good point about the backflow EGT01! -sshep

If the recycle valve fails open then there should not be any pressurisation of the compressor suction. If the compressor is considered as a black box then taking all the discharge to the suction stops the flow of gas into and out of the box but does not increase the contents (and therefore pressure) in the box.

The risk of backflow through the recycle from the discharge system into the suction system should be considered during the design and suitable precautions taken to prevent problems in the lower pressure system - installation of NRV, slam-shut valve etc.

Suction drums should be rated for settle-out pressure when the compressor is stopped.