The water heats up, building pressure, and then they vent some steam to try and lower the pressure (and then, obviously, have to re-add water). The explosions h

The water heats up, building pressure, and then they vent some steam to try and lower the pressure (and then, obviously, have to re-add water). The explosions happen because the steam separates into Hydrogen & Oxygen due to a reaction, and once outside that is an explosive mix. The chamber is pretty large, several meters deep around the core. The explosions are as a result of the release of the hydrogen/oxygen mix - which does not take much to catch - but they occurred outside the pressure vessel. Such a situations is technically planned for, the building structure "fails outward" allowing the blast out and away from the pressure containment. It is relatively unlikely to damage the integrity of the vessel in such a situation; they are very very strong pieces of kit (at three mile island the explosion took place inside the vessel and I believe it did not compromise its integrity). At this stage is somewhat unclear but it appears that the third explosion has damaged one of the pressure vessels. However the technical explanations I am seeing are stating that it hasn't compromised the integrity of the containment; instead it appears to be saying that parts of the mechanism inside the containment are damaged (in this case the "torus" which is used as an emergency pressure relief system). This is actually, though, a significant problem, if true, because it could affect what happens if it finally does meltdown (as I mention in another comment, my working knowledge of this form of reactor is incomplete, but IIRC the torus also forms part of the final emergency response to a meltdown). Bottom line is; at this stage we know there have been 3 hydrogen explosions at 3 of the units, and one fire at another of the reactors. But it is still unclear the level of damage and risk with each unit - TEPCO (and indeed Japanese companies in general) tend to play down or not discuss details and risk publicly. Take this fire; no one really picked it up until it was out - at which point TEPCO appear to have started mentioning it. If I had to lay a claim I suspect that reactor 2 at this stage will probably meltdown; loss of the torus is significant enough that it raises the risk of what they are currently doing (i.e. trying to cool the core).