Then once the IP address has been assigned, the ADC function would check the firmware of the replacement drive and if necessary upgrade/downgrade to the proper version...
Just a small nitpick - the ADC doesn't do firmware. For that, you need firmware supervisor which is a completely different setting.
In my opinion, ADC without firmware supervisor is nearly useless, because I can almost promise that whatever drive you get off the shelf 5-10 years later to replace the failed one will have a different version of firmware. I never use one without the other.
To enable firmware supervisor, you need to set the compatibility setting for all the relevant drives to "exact match" (instead of "compatible module", which is the default). Then you need to load the firmware files onto the PLC's SD card by going to controller properties>non-volatile memory. You will have to switch the PLC to program mode while you do this. You must have the desired firmware files your your PC to be able to download them to the PLC's SD card, of course.
To give my answer to the OP's original question - I use ADC and firmware supervisor wherever possible. I have never yet set up an application with DHCP persistence because the only one of my clients that had that type of network structure, only had support for DHCP and not BOOTP. Powerflex 525 drives only support BOOTP unless you add the additional comms module.
My electrical drawings always contain a page (or pages) for parameter lists for all parameterisable devices, so when I'm using this strategy, my parameter list page for the VSD's consists of a short brief on what ADC/firmware supervisor is, and how to replace a failed drive - i.e. get the IP address/subnet mask/gateway from drawing ABC, enter them into parameters XYZ, and let the PLC do the rest.