I was referring to the question of how to stop chain-reactions in dedicated parking garages full of cars.
While a chain reaction can easily happen
inside a battery pack, the risk of it happening between two battery packs at least a meter away from each other, is extremely small.
The energy release of a lithium ion fire is, ultimately, not
that massive (I remember seeing analysis that when this happens with packaged cells, the energy release is the same order of magnitude than the cardboard packaging material combusting). The order of magnitude we are talking about is something like 100-200kWh released within 5-15 minutes, equivalent to a good big bonfire. An equivalent gasoline fire would be around 500kWh (for 50 liters or 13.2 gallons) possibly released within a minute or even less; yet, AFAIK it's rare that gasoline cars pop up in a chain reaction, even though gasoline car fires are extremely common and happen all the time (which is why, btw., all the big public parking garages have sprinklers, at least here... I think it's mandatory). Actually random gasoline car fires are so common they are not typically reported as single cases by the media, they go into statistics only.
Now, the issue
within the pack is that the neighbor cell start to run away at around 160 degC. The problem: the cells are just
millimeters from each other - or completely in contact, with no thermally protective material - and in a big pack, it's a dense 2D matrix, so once you have a few cells going, they heat up the neighbors from multiple sides. The heat has nowhere to go, except to the other cells.
For example, IIRC, Tesla has claimed they have analyzed their pack can withstand any single cell running away - this makes sense as they have small separations (5mm? Not much, but more than usual) between the cells, and cooling pipes running between them. So, the fires have been caused by, for example, large enough physical penetrations to put multiple neighboring cells into thermal runaway.
In any case, this gives you the
idea how close the heat source needs to be in order to cause the damage. Conceptually, you can simulate this with candles: when you really lump them together in a large matrix, there is a reported risk of a chain reaction, but once you separate them by
an inch, that problem is gone.
Cars tend to be separated by at least half a meter, and the battery packs tend to be enclosed in metal, and don't tend to protrude near the car's external measurements. So, the pack-to-pack distance would be around one meter a least, and there would be multiple metal walls between, spreading the heat. Comparatively, take a modern fire stove; it has double steel wall with an outer envelope, with air cooling channels inbetween. I can safely tape an 18650 cell to such a stove, despite the >1000 degC fire going on inside, just 10 cm away!
While a burning car might set the neigbor's wheels on fire, no electric vehicle should have such a massive design failure that the wheel fire spreads to the battery pack (the opposite can happen).
And, when the fire is in open place, convection quickly cools and mixes the air, moving the heat to a larger area. If there is any hotspotting, it tends to be right
above the fire. So maybe with tightly stacked cars without solid floor inbetween, as in a large multi-car tow truck / car transport, this could be a problem?
Within separate cars with enclosed battery packs, it would require
very strange circumstances to bring any cell over its runaway temperature. Might be possible with an unfortunate vent hole directed
right at the worst direction where a poorly protected battery pack of the another car sits, but cars seldom are parked within centimeters of each other, so the flame would need to somehow be extremely focused, and basically direct the full combustion energy of the pack to one place, in a narrow beam that exactly hits the battery pack of the next car. And for this to become a massive chain event of more than two cars, this unfortunate condition would need to repeat with every car in the line.