Isolation is usually an absolute requirement, silly poet. ;)

Integrated charging (using the regenerative braking circuitry to charge the battery from the grid) is rarely used despite its enormous cost advantages, precisely because it lacks galvanic isolation. The double insulation, isolation monitoring, and other techniques such as performing an emergency power down if someone opens the hood while DC high voltage is active do not satisfy most specifications. I think this is likely to change, someday. But there are other complications to integrated charging such as poor efficiency at low power (such as 120 VAC source), multiple optimizations required (surviving power grid transients during electrical storms is quite different than regenerative braking), etc. I am guessing here but you get the idea.

See the Renault Zoe as an example of integrated charging (43 kW with a cheap 3 phase AC EVSE!) and the BMW Mini E of years past (AC Propulsion drive train, still almost 50 cars running in Delaware in V2G power trials).

EP, surely you heard that Tesla dropped the AC Propulsion drivetrain due to reliability problems? Surge suppression is not enough apparently. My personal experience with AC Propulsion in the Mini E was roughly a charging system failure about every 6000 miles, a total of 4 such failures over 11 months. Maybe you can fix it for them with surge suppressors, good luck. ACP is still selling many hundreds of drivetrains per year from what I have heard, maybe you should send them your resume? I have to admit that car was an absolute hoot, as much trouble as it was. Perhaps the main issue was the older analog controls, I dunno.

Still I agree that things will probably go towards integrated charging in the long run, especially for heavy duty applications. But it is not hard to see why most everyone (including Tesla) is shying away from it for now. I would not be surprised to see it come back in the Model 3. That would be cool.

And the isolation issue is about the need to remain safe even after a single failure point, you do understand that requirement? I have the impression that permeates all of the automotive world as well as the National Electrical Code world.

But certainly this article is about the conventional isolated charger design with a small high power high frequency transformer. As I read it, they are talking about converting the incoming 60 Hz (or 50 Hz) AC directly into high frequency AC to feed the isolation transformer, without rectifying it first. Sounds neat, what is wrong with that if your project management insists on galvanic isolation?

I think Ideal Power has an interesting approach to isolation without a transformer with their time division technique, they call it Power Packet Switching. But the questions of isolation after a single point failure remain. Transformers are pretty good about isolation, it will take time to build confidence in alternatives.

EP, ground fault interruption is a mitigation technique, not a prevention method. At the voltages commonly used in EVs, a GFI (called an RCD in Europe, residual current detector) might not prevent a death in the event of a fault, it may only prevent the corpse from being burned beyond recognition. Bad joke, gallows humor and all. But the point is that prevention is a requirement, mitigation is not enough.

I don't know what the reliability problems were with the ACP integrated charging method. I hear that the Renault Zoe sales are taking off, and the Zoe uses essentially the same techniques, as far as I can tell. (The ACP patents are long expired.) So I presume Renault has fixed the issues.