Machingbird (I like that one):
Basically the same effect as substituting a lesser capacity battery cell in series with the remainder of the battery cells. No way to avoid abusing the 'smaller' cell with the original design if that happened.
Has this been discussed elsewhere?
Have not seen discussion, other than the argument the the "cell" should be treated and protected as if it were three cells in parallel, or physically or virtually isolated and should be monitored and protected as such. But if the kind of failure you propose happened the result would depend in part on how much adjustment is possible within the battery balancing network. If it is working properly, that cell will not be driven below LVCO because the BMU will tell the charger that it had a problem and the battery itself would open the contractor within the battery if the load or charging did not stop.
Nor will the remaining two sub-cells be forced to take current once their voltage reaches the upper limit. That would have the effect of limiting the whole batteries performance to the low to high capacity range of that partially failed cell. The resistive heating would be slightly larger, but not enough for an thermal cascade or venting.
But that assumes that the balancer wires and pass element for each cell can carry the full ~40 amp charging current, and that is very unlikely given the size of the wires shown.
You may have a good point, and the CAT scans of the APU battery showed some internal collector wire to electrode disconnects within several of the cells. But that was assumed to be a result rather than the cause of the failure. Some of the 1/3 cell rolls did look different from their neighbors.