Momoe;
I'm still slightly surprised that there were no indications of potential battery problems during the testing/flight testing regime, these normally take each system to the edge of the envelope, yet no issues reported?
Not so close to the edge during flight tests. Subsystems are usually pushed much harder during bench (lab) testing. And therein lies a possible problem. With much of the testing being done by the subsystem vendors, the interactions between different subsystems may not have been tested adequately.
Its more likely* that Thales loaded the battery system down with a resistive load bank to simulate an APU start rather than actually obtaining production DC-AC converters and an APU starter/generator.
*Disclaimer: I wasn't there. But this has been my experience with past subsystem tests.