I presume the cooling will be a bit easier than than the precooler on the SABRE rocket engine: https://en.wikipedia.org/wiki/SABRE_(rocket_engine)
If cooling is still an issue due to size then I presume running multiple DPFs in series might be an option?
More like using a genetic algorithm to control the change in axial field based on some monitored output, which would then find the ideal variance in axial field to produce the best energy gains.
I would expect there to multiple DPFs per plant which can be ramped up during peak demand, then individual DPFs can be maintained during times of low demand.
