Harm Otten had a team from the local electricity distribution company over for an unrelated metering problem, and obtained from them a screen shot of the voltage / current curve of ZOE charging. This is a Q model.
The dark red voltage measurement is between one phase and N. The current, light red line, is measured in that particular phase wire. The current flat line around 0 is consistent with rectifying, and the “bump” after the current has swung up is consistent with 3 phase rectifying using a three-phase full-wave rectifier (see also this post).
Note: it all makes far more sense now, see this post, ignore what I wrote below.
What is harder to understand is why the charger is not able to let the current curve follow the voltage curve better, given the actual design seems to have full input control (see Charger design post).
Edit: I happen to believe, though I was not there, that they used a Fluke 435 using Fluke i430 current probes (Rogowski coils) for current measurement. That is pretty top of the line equipment, so unless they didn’t match up the voltage and current phases, I cannot explain the rather massive phase shift.