It's laid out pretty clearly in http://mitnse.com/2011/03/13/why-i-am-not-worried-about-japans-nuclear-reactors/?mit http://mitnse.com/2011/03/13/why-i-am-not-

It's laid out pretty clearly in http://mitnse.com/2011/03/13/why-i-am-not-worried-about-japans-nuclear-reactors/?mit http://mitnse.com/2011/03/13/why-i-am-not-worried-about-japa... which was posted to HN today as well. In short, even after reactor shutdown, substantial heat is still produced, so some cooling is needed. Substantial as in, up to some 7% of normal output, early after the shutdown. That's mostly due to residue fissible elements -- byproduct of normal operation. The explosion was not inside reactor pressure vessel, but in the housing. Top parts of fuel rods got exposed out of coolant water at some point and heated up considerably. When water level was raised back again, heat caused release hydrogen. Upon subsequent steam venting (necessary in case of overheating), the hydrogen combusted in the air in reactor housing (i.e., building, not the pressure vessel itself). The pressure vessel itself contains no air to avoid explosion. Compare a very early mishap with a nuclear pile (not reactor): http://en.wikipedia.org/wiki/Windscale_fire http://en.wikipedia.org/wiki/Windscale_fire -- where the Brits tried to extinguish smoldering graphite & very hot fuel with water, and there was risk of hydrogen explosion (fortunately avoided).