@edyong209 tweeted that this was the “Best explanation I’ve read of how nuclear power plants work”

Overall it’s not too bad; I’ve seen worse, and there is some good information. But let’s look at what the reporter got wrong.

A fission reaction is a lot like a table filled with Jenga games, each stack of blocks standing close to another stack. Pull out the right block, and one Jenga stack will fall. As it does, it collapses into the surrounding stacks. As those stacks tumble, they crash into others. Nuclear fission works the same way–one unstable atom breaks apart, throwing off pieces of itself, which crash into nearby atoms and cause those to break apart, too.

I’ll ignore the unquoted part where he treats heat as a substance (a far more ingrained conceptual issue). The atom throwing off pieces of itself is really throwing off two fission fragments, which don’t go very far — they’re highly charged nuclei (the electrons get left behind) and they deposit their kinetic energy in a short distance, which is where most of the energy is deposited, and why the reactor heats up. The parts that cause more fissions are neutrons. They are uncharged, and can travel a greater distance — they don’t have to hit an adjacent nucleus. The Jenga analogy isn’t horrible, but it’s not great, either.

The neutrons don’t cause another fission because they have lots of energy, which is implied by the description. Quite the opposite — a slow moving neutron has a greater chance of interacting with a U-235 nucleus and inducing fission, which is why you put a moderator in the reactor — it’s something the neutrons can hit and lose energy to, but isn’t likely to capture the neutron — you don’t want to lose any more neutrons to non-fission reactions than you have to.

The author continues to imply that the atoms hit each other through the article; I’m not going to call out each instance.

Proximity is also what makes the difference between a nuclear bomb, and the controlled fission reaction in a power plant. In the bomb, the reactions happen—and the energy is released—very quickly. In the power plant, that process is slowed down by control rods. These work like putting a piece of cardboard between two Jenga towers. The first tower falls, but it hits a barrier instead of the next tower. Of all the atoms that could be split, only a few are allowed to actually do it. And, instead of an explosion, you end up with a manageable amount of heat energy, which can be used to boil water.

Control rods aren’t the only difference between a bomb and a reactor. If you somehow managed to pull all the control rods out of the core you’d have a nasty nuclear accident on your hands, but no nuclear explosion.

When a reactor core is shut down, its energy output drops not to zero, but about 6% of its normal output, Forsberg told me. The reactions grind to a halt over the next few days, as the falling Jenga towers run out of other towers they can actually hit. In the meantime, atoms keep breaking apart, releasing both heat and fast-moving particles that can penetrate human skin and damage our cells. Because of this, every nuclear reactor has ways of getting rid of the heat, and blocking those fast-moving radioactive particles.

and at the end

And then what happens? Remember, this is really just an emergency shutdown gone awry. The control rods are still in place. The Jenga columns are still separated. So, over time, the fission reactions will still slow down and stop. As they do, heat levels will drop, and so will levels of radiation.

The author first implies and then explicitly states that fission reactions are the cause of this power output, and it’s not. As I previously explained, this power comes from the decay of fission products. The reactions slow down because the short-lived products decay away quickly (which is why they are called short-lived). Not fission — this fission rate has been reduced by many orders of magnitude, to the point where heating from it is negligible.

Edit: You really should read this.

I’m not an expert on nuclear reactors. I taught in the nuclear power program of the US Navy some years ago, meaning I was competent to discuss some aspects of reactors, and specifically the type the navy uses. So I’m also not some random guy in the street. With that disclaimer in mind, there are a few items to mention with regard to the reactor issues in Japan following the earthquake.

This is not another Chernobyl. The reactor design is very different, and the circumstances are different. The Chernobyl accident (link for the more technically inclined) involved an operating reactor that went prompt critical as the result of operational errors, deliberate disabling of certain safeguards as part of a test, and design flaws. This caused a steam explosion and chemical fires as the carbon moderator caught fire.

A closer analogy would be Three Mile Island.

There have been reports of an explosion, but it must be stressed that this was not a nuclear explosion. The reactors have been shut down. It’s not so easy to cause a nuclear explosion in the first place (bombs require a level of expertise), and a shutdown reactor does not have the capability of sustaining the fission reaction. This leaves us with steam pressure buildup or hydrogen as the most likely culprits, i.e. it’s thermodynamics or chemistry, not nuclear physics, which explains the explosion.

The reactor is shut down, so what’s the danger? The products of a fission reaction are typically radioactive, and subsequent decays also release energy. Shutting down the reactor reduces the fission rate by many orders of magnitude, so it’s effectively zero in terms of heat output, but the radioactive fission products still release up to 6-7% of the plant’s power output. The actual value depends on the operating history; the fission products with long half-lives take longer to build up to steady-state values. This value will drop fairly quickly as the short-lived isotopes decay, but it’s still significant — a reactor rated at 1000 MW will still be producing tens of MW of decay heat. The reactors in question at Fukushima Daiichi are rated at 460 or 784 MW (edit 3/15: AFAIK that’s electrical output; if so, the thermal output is ~ 3x higher)

So shutting down does not mean it’s Miller Time? Right. You need to run pumps and do something with the energy, which usually means piping water to a cooling tower, which means you need to run pumps, and those require electricity. It seems silly, at first glance, that a reactor would need a source of power to run it, but the turbines are probably designed to run at the high power output of the reactor and not off of decay heat. So you have an external power line (lost in the quake), local generators (apparently also damaged) and battery backup. Redundant systems. However, it seems that the damage was severe, so the primary and first backup systems are still offline, and if cooling was lost (batteries have a finite lifetime), the water in the core can boil away.

That sounds bad. Yes. As long as the core stays covered with water, things should be fine. But uncovered, the temperature can rise and fuel elements can begin to melt. Hydrogen is produced, which can explode, and boiling water becomes steam, which raises the pressure in the containment vessel. The latter is why the containment vessel would be vented. You would need to replace that water into the system, which also requires pumps. (This what had happened at TMI, though in that case, the cooling pumps were shut off deliberately owing to a flawed procedure)

So this is serious. Nothing here is meant to imply otherwise. But the term “meltdown” (or worse, if preceded by “Chernobyl-like”) raises all sorts of imagery, most of which is inaccurate.

Here are some links from what look to be credible sources. This is a dynamic situation, so there is a shelf-life to the details.
Nuclear Crisis in Japan: What We Know
Factbox: What happens when a reactor loses coolant