It’s just great how you illuminate this issue.
To transfer externally 20A after is the better solution. Does not affect the protection of the aluminum box.
But your Fuse Holders are not responsible for AC100 … 240V!
In addition, protection against water and dust important (IP54 … 65), and shock protection by fuse replacement.
Perhaps stuck in heat shrink tubing?

Right, why 20A blow, we have to accept unexplained for now.
20A fuse is typ “fast blow”_ AC / DC converter could inrush_ … etc.

My house fuse is a miniature circuit breaker L-16A(for Line Protection). These have magnetic and thermal releases combined (Does USA also determined.).
The characteristic curves have an intersection at about 6 times the rated current. It is triggered because of 96A in about 5 seconds. At 1.4x the rated current no tripping.
So is neither protective nor of importance for the 20A fuse in the PCU.
I must go to try to measure over an hour the current profile for AC230V charge Inlet. Where I previously said it would under 14A.
Merci HANS

I agree that there is a fault somewhere in the PCU/charging system or possibly in the EVSEs or grids attached to the failed cars or maybe even in the traction batteries. The intention is not to install a 25A fuse to accommodate proper operation. 25A is a somewhat arbitrary value for the fuse just as long as it’s rated higher than the inline fuse. The intent is that it never blow and that the inline 20A fuse maintain the specification that the manufacturer designed.

If you are confident that your AC line never flows more than 14A while operation normally I would highly recommend installing an 18A fuse inline (I use the example 18A because I have seen them available as I shopped online for fuses; 15A would work, too, of course.) This would guarantee that your easily installed/replaced 18A fuses blow before the 20A fuses that are very painful to replace.
If we can “prove” that every Think on every charger will not exceed 14A then I would recommend installing the inline fuse for everyone without modifying the fuse inside the PCU. My design simply attempts to avoid any situation where the inline fuse might blow while the system is operating properly. I know that some of the Think owners in my geographic area monitor their charging closely. I’d like to collect more data to make sure that there aren’t peak charging currents that exceed the average before recommending it to my customers, regardless of what the specification says.
As far as isolation faults, I have seen coolant leaks into PCUs and a water leak into the traction battery that caused isolation faults, but none of the PCUs with blown charger fuses had any moisture intrusion in the PCU. I don’t think water in battery would have caused it, especially without a flashing high voltage warning light, but I would never gamble on what I think.
AC charge source properties: Surely EVSEs include a surge protector, right? I don’t know, maybe not… Perhaps I should go back correlate the EVSE brands to the failed chargers. (I suspected that we already noticed that at least a few different EVSE brands and stopped taking note.)

I hope I’ve answered your questions– I’ve done my best. Your input is very valuable; I really appreciate it.