Originally Posted by saptzae
@Machinbird
3-4s of high current (45A) caused the conflagration! I estimate that BMS has less than 100ms to prevent electrically induced runaway after a cell short.
Thus am I afraid that temperature sensing will be too slow. Cell pressure sensing may be an interesting option.
saptzae
Referring to the Boston battery failure,No one knows what temperature the cell that initially failed was at just prior to failure. The fact that the short appears to have occurred on the centerline of the plates is indicative of a temperature factor in the failure. If an overheated cell is detected early in the charge sequence (because the sensor is closely coupled to the cell), then current can be attenuated to that cell or the charge sequence aborted.
Once an actual failure begins to occur, then a pressure sensor or a rapidly responding current sensor can abort the charge and flag a battery malfunction, but that is after a failure. The main idea should be to never let the battery (as a whole and any portion of it) reach any of its limits.
My personal approach would be to charge the battery with a series of cell voltage chargers, each one floating at the potential of the actual cell it was charging.
Yes you would have more wires running into the battery, and you would have cell sensors running to the appropriate chargers, but there would be no need of a Battery monitoring unit in the battery itself. The chargers could record and control the status of each cell. If one cell required twice the time to accept a charge as its neighbor did while staying within temperature limits that fact could be accommodated and reported if appropriate.