1021:01 APU battery bus voltage decreased from 32 to 31 volts.
1021:04 APU battery current increased to between 44 and 45 amperes for 4 seconds, indicating current flowing into the battery.
My reading of this is that the battery might have been charging when the voltage suddenly dropped a bit. If the voltage dropped due to a partial short circuit in a cell that would probably result in the charge current increasing.
Initially the voltage only falls 1V which is probably not enough for the BMU to see it as a fault (The battery voltage might vary that much due to normal load variation?). That might explain why it continues charging for another few seconds. Only when the voltage has fallen far enough does it realise there is a battery fault and issue the EICAS message.
The way the voltage fall sounds like one cell in trouble because it's dropped roughly one cell voltage (eg 4V).
The strange thing is that the system knew the battery had failed (EICAS message) but the contactor hadn't been commanded to disconnect the battery from the aircraft.
I notice the battery voltage falls to 0V and returns to 28V. I wonder if that was caused by the contactor trying to disconnect the battery? I hope the board controlling the contactor doesn't need a functioning battery to provide power to control the contactor?
Edit: It would be interesting to replace a cell with a box of electronics that simulates a cell. See what happens if you just slowly turn down the voltage of this dummy "cell".