Japan Nuke Design: What they did right and what they did wrong

Why yet another article on the nuclear disaster in Japan? Because there are a lot of partial truths and untruths by the MSM. It's only by piecing them together and resolving the contradictions that a truer picture emerges.

Disclaimer

I am not connected in any way with the nuclear industry. I have no formal training in nuclear power generation, although whilst gaining my degree in Electrical and Electronic Engineering I was taught about conventional (coal/oil) generation (steam is steam and generators are generators, only the heat source changes). I am not defending or condemning nuclear power here, merely considering the engineering decisions that were made, whether they were good ones or bad ones, and the implications.

Background

The Japanese nuclear plants affected by the tsunami were essentially General Electric Mk 1 Boiling Water Reactors with a few modifications. GE Mk 1 reactors are still in use in the US.

What They Got Right

To be more accurate, what they got right together with what they didn't get wrong (decisions that were not great but not wrong either).

The decision to use nuclear power at all

Given the shortage of free land in Japan, its economic circumstances and—above all—the assurances of the nuclear industry of safety, this wasn't an obviously wrong decision at the time. It still isn't an obviously wrong decision, but the implementation clearly left a lot to be desired.
Fitting vent valves

The original GE design did not have vent valves for the core. That meant that a loss of coolant accident (LOCA) partially exposing the fuel rods could generate a lot of hydrogen and oxygen eventually leading to pressures that could crack the core open even if they didn't explode. The designers claimed this could never happen because of all the redundant safety systems.

After Chernobyl and Three Mile Island, countries around the world started modifying their Mk 1s to have vents. Japan resisted doing so until 1992 but finally added vents. These were used in the early stages of the incident to bleed off hydrogen and oxygen.
Siting plants on the coast

Any thermo-electric system, be it coal, oil or nuclear, requires both a supply of heat and a sink of heat (i.e., something cold compared to the heat source). The thermodynamic efficiency increases with increasing temperature differential of the two. Properties of materials place an upper limit on the temperature of the heat source so good cooling is essential for high efficiency. The heat sink must not only be cold, it must be able to dissipate the heat without itself getting significantly warmer.

One method of cooling is the cooling tower. Immense structures shaped like fattened diabolo toys lazily emitting water vapour—they take up a lot of space which is short in supply in Japan. Fast-flowing, high volume rivers are good for cooling, but they're in short supply. Many nuclear reactors in many countries are sited on coasts to use the sea as a heat sink.
Quake-proofing

The reactors survived the quake. Despite the fact that the quake was of higher magnitude than the worst-case the designers expected. Despite the fact that the underlying rock was a lot more friable than it was believed to be (it was claimed by TEPCO to be solid but you could crumble it in your hand). It is possible to design reactors to withstand earthquakes and these did withstand a more severe earthquake than anticipated, despite the current outrage about the stupidity of siting reactors in an earthquake zone.
Three backup generators

The GE Mk 1 design (as used in the US) had two diesel backup generators in case power from the grid was lost. One of these could provide sufficient power to pump cooling water into the reactor core if there were a LOCA. If that generator somehow failed, the second was there as a backup.

The Japanese reactors had three backup generators. Even if one were in pieces due to maintenance/repairs, there were still two generators available.

What They Got Wrong

These are seriously flawed design decisions.

Choosing the GE Mk 1

The Mk 1 was cheaper than competing designs because, amongst other things, the containment was not as strong. It has attracted criticism over the years, including from regulatory authorities, but no action was taken.
No vents in the reactor building

After retro-fitting valves to the reactor core to vent off hydrogen and oxygen in the event of a LOCA it appears they gave no consideration as to the consequences of venting. According to reports the gases vented into the main building rather than being vented to the outside. This appears to be the cause of the explosion that blew the cladding off the reactor buildings leaving only a skeletal steel framework, the reactor and associated machinery.

At best this was a stupid design decision but not a catastrophic one. At worst was a catastrophic decision as it could have destroyed internal power lines and control systems that would have taken out backup generators and/or pumping systems necessary to prevent meltdown.
No allowance for tsunamis

The word "tsunami" is Japanese and has entered into usage worldwide because tsunamis are frequent occurrences in Japan. It is beyond belief that the Japanese reactors did not take tsunamis into account. Yet the emergency diesel generators were sited close enough to ground level that the tsunami flooded all three with water and particulates, rendering them inoperational. This is design stupidity bordering upon criminal negligence.
Placement of "cooling pond"

A new nuclear power rod, in isolation, has so little radioactivity and so little heat that you can safely hold it in your hand. The size and geometry mean that there is a very low level of spontaneous fission but very little induced fission by neutron capture.

A used power rod, however, is very hot in terms of both radioactivity and temperature. Radioactive enough that it is dangerous to go near it. Hot enough that without cooling the cladding, and then the fuel itself, would burst into flame releasing highly-radioactive particles into the atmosphere.

For these reasons, spent rods are placed into cooling ponds for years until the radioactivity declines enough that they can be removed. The water of the pond keeps them cool and attenuates radioactivity to safe levels. A cooling pond that loses enough water to even partially uncover the rods results in a serious fire releasing very radioactive and toxic material as fine particles that can be carried by the wind.

In the GE Mk 1 reactor, the cooling pond is not a separate facility. Instead it is a tank in the reactor building. It is separated from the reactor core by around 1/4 the diameter of the core (visual estimate from various graphics). The vertical centre of the pond is horizontally aligned with the top of the core (again, a visual estimate). A serious hydrogen explosion in the core would almost certainly damage the pond too (as seems to have happened in at least one Japanese reactor).

The placement of the pond makes transfer of spent fuel rods from core to pond far simpler, safer and cheaper. But it seriously magnifies the dangers of a core explosion both in terms of quantity of radioactive and toxic materials released but in the intensity of the radiation. Another design flaw bordering upon criminal negligence.

Conclusions

The GE Mk 1 reactor has intrinsic design flaws whether it is sited in an earthquake/tsunami zone or not. Of interest is how these design flaws came to be. One GE scientist quit because of them. Nuclear regulatory agencies have spotted them over they years and done nothing about them. Why?

GE marketed this reactor as "intrinsically safe," saying that failure modes would not cause harm either because the failure mode was self-limiting or because of multiple backup and emergency systems. Was this due to inadequate analysis? Certainly the lack of vent valves in the initial design were because they were deemed unnecessary as LOCAs would be dealt with by backup systems.

I doubt GE had incompetent engineers on such a project. Part of the problem would have been "bean counters" insisting that cost be minimized so they would get more buyers. Part of the problem would have been senior management with more financial than engineering expertise insisting that, for example, vent valves not be fitted because it might be interpreted as an admission that perhaps the reactor wasn't as intrinsically as safe as was claimed (good engineers would see vent valves as "belt and braces" just in case rather than as a criticism of other parts of the system).

Nuclear regulators dismissed the flaws because admitting to them might cast doubt on the whole concept of nuclear power and/or impact the profits of GE. Never underestimate the power of politics and money to drive bad decisions.

Given the forces that permitted the design flaws in the GE Mk 1, we have to seriously question if any existing nuclear reactor design is safe, do a thorough analysis of them all, then take the appropriate action. It may not be economically feasible to rectify flaws in some designs and therefore it may be necessary to shut down those designs as soon as possible.

We must not approve any new reactor designs or even new reactors to old designs until the regulatory process is fixed. No more "head in the sand" because ordering changes might cast a bad light on nuclear power in general, or cost GE or plant operators a lot of money. The regulators must place safety above all. I am not confident, given the influence of wealth and politics upon such organizations, that this is possible.

There are a couple of interesting reactor designs that might be intrinsically safe. Our continued lack of action towards the twin problems of Peak Oil and Global Climate Change may mean we have no choice but to use nuclear power to supplement renewable energy sources when the excrement hits the fan in a few years. Our procrastination means that when things get bad we may not be able to deploy sufficient renewable energy sources quickly enough. If our lack of action now forces us to temporarily supplement green energy with nuclear power we need the safest technologies available. The two possible candidates I know of are the Pebble Bed Reactor and the Travelling Wave Reactor—but in light of recent events I take the claims of both with a large pinch of salt.