Hacker News Reader

Comment by koski | Hacker News Reader

what I understand now is that the blast at the nuclear plant may have been caused by a hydrogen explosion (pressure) or If nuclear fuel rods overheat and then come into contact with water, this produces a large amount of highly-flammable hydrogen gas which can then ignite. Right?

Can anyone with more knowledge and experience of this confirm?

View on HN

That's certainly possible, theoretically. However, I'm not sure if those conditions can actually be met (rods should not ever come into contact with water and nothing should be able to ignite the gas, especially not at higher pressure levels).

No, actually the fuel pellets never need to come in contact with water; it is their cladding, which is a zirconium alloy ("Zircaloy"), which can react with steam under extreme conditions. It's a redox reaction: Zr + 2H2O --> ZrO2 + 2H2 or something similar.

This actually happened at Three Mile Island, and there was a hydrogen explosion in the containment building there (which did not break).

(I am not an engineer)

That makes more sense, thanks.

The explosion was likely a containment failure of the previously reported (radioactive) steam that had been building up in the containment building.

More info: http://www.world-nuclear-news.org/RS_Battle_to_stabilise_ear...

An attempt to explain the risk to the Fukushima nuclear plants following the earthquake: The plants are designed to shut automatically, which halts the main nuclear fission reaction, but there is a residual amou ntense heat within the system. Back-up generators should kick in to power the coolin mechanisms needed to dissipate that heat - b they fail, as appears to have happened here, temperatures rise. If this isn't stopped, the reactor vessel itself could eventually melt an eak.

Right? Has anybody found any documents if Japan has been notified or warned about these backup systems of these plants? It sound like a design failure to me. (I know nothing about nuclear power plants)

Reply on news.ycombinator.com