Despite having three degrees related to nuclear engineering, power generation and nuclear physics, I’ve never actually worked in the nuclear industry, as it became completely moribund around the time I started my Master’s, which more-or-less coincided with the explosion, fire and meltdown at Chernobyl. My post-doctoral experience focused on radiation detection systems and radiation transport calculations, both in neutrino physics and medical physics.
As such, I’ve been asked by a number of people over the last week about the current problems in Japan, and intend here to pull together some of my thoughts and answers to those questions. For anyone who wants a simplistic, “Yay nuclear power is great” or “Nay nuclear power is horrible”, I suggest the works of Petr Beckman and Helen Caldicott, respectively, although neither has anything useful to contribute to the genuine and serious debate we need to have regarding the future of nuclear power and whether it has one in our both literally and figuratively power-hungry society.
Fukushima Reactor Design
To understand what has been going on a Fukushima a look at the reactor design is a good place to start. Wikipedia has some good information, including this image:
The reactor core is contained in the yellow pressure vessel. The core consists of fuel rods and control rods and their support structure. This style of reactor is based on power plants for nuclear submarines, and is no longer considered quite the done thing, for various reasons I’ll get to below.
For now, notice that outside the reactor pressure vessel we have a steel containment vessel (the orange light-bulb) connected to a torroidal water-filled pressure-relief vessel, and surrounded by a concrete secondary containment structure. Unlike the Soviet Socialist triumph that was Chernobyl–which was basically a carbon-moderated core sitting naked in a heavy warehouse–this reactor has three layers of protection between the core and the workers. The building structure constitutes a fourth layer of protection for the public, albeit not a very useful one.
There are any number of interesting passive safety features in this design, although my eye is drawn particularly to the concrete well below the reactor pressure vessel (the open space where the numbers 31, 13 and 19 are). The supporting members may be just pillars, but I can’t help but wonder if there isn’t a full annular wall, forming a well to catch the melt should the worst happen. It would certainly ease clean-up to have things in a nice compact geometry like that. But that is speculation based on insufficient information, and however fun to indulge in doesn’t create knowledge, only avenues for further inquiry.
A less exciting feature of the design is the placement of the control rod actuators beneath the core, so the rods are pushed upward to shut the reactor down. Gravity is a weak but remarkably reliable force, and in the event of an accident I would prefer to have it on my side rather than be working against it.
The least exciting feature of the design is labelled “5″, which contains “27″. This is the now-famous spent-fuel storage area, about which much is bound to be written in the coming weeks and months.
While the reactor core has no less than three barriers between it and the outside world, the spent fuel has basically nothing, and in the unfortunate event of a loss of cooling water has the potential to melt down from residual heat, depending on how old it is.
Nuclear Fission Made Simple-ish
Fission reactors are simple in principle but complex in practice. Fissionable nuclei, like those of some uranium, thorium and plutonium isotopes, are like large wobbly drops of water, which can be split apart (fissioned) by being hit by a slow-moving neutron. The fission process itself creates more neutrons which can be slowed down by a moderator, and these neutrons can cause other nuclei to split, thus creating a self-sustaining chain reaction so long as you have enough material in one place, and there is nothing around to absorb all the neutrons.
Neutrons are like ghostly quantum tennis-balls that pass through matter easily until they happen to hit another nucleus head on, and then they scatter off it in some other direction, losing energy in the process. Typical fission neutrons can travel a metre or so while slowing down, and this sets the dimensions of the typical reactor core. We will never have fission reactors with dimensions of much under one meter.
As well as scattering off other nuclei, neutrons can also be absorbed by them, changing their atomic mass and in most cases creating a radioactive isotope. This happens a lot in the general vicinity of fission reactor cores, where the neutron flux is extremely high, and results in pretty much everything inside the containment being significantly radioactive after a while.
If you go back and look at the picture, you’ll see at the top of the reactor pressure vessel there are heat exchangers. These are made of various exotic alloys, and unlike the heat exchanges in coal-fired plants, they are quite radioactive.
The reason for exotic alloys in nuclear plant construction is that neutron radiation tends to damage the crystalline structure of pretty much everything. For example, about 6% of iron has an atomic mass of 54 (26 protons, 28 neutrons). When iron-54 absorbs a neutron, it becomes iron-55, which is a radioactive isotope that decays by capturing one of its atomic electrons to turn a proton into a neutron, which creates manganese-55, the sole stable isotope of the element adjacent to iron in the periodic table. So now what used to be an iron atom sitting in in a micro-crystal of iron in a steel tube is now a manganese atom. Manganese has different physical and chemical properties than iron, and as time goes on this kind of thing happens to more and more atoms in any material in the reactor. This results in extensive micro-crystalline damage that is pretty much exactly what you would do if you wanted to induce corrosion and cracking as rapidly and extensively as possible.
My understanding (although I hasten to point out that I am not a chemist!) is that corrosion in this environment is proceeding by something like:
H2O + m’ -> mO + H2
where m’ is a metallic atom in some awkward state in the lattice–like a manganese where an iron should be–and mO is the metal oxide that gets formed, freeing the hydrogen.
This sort of chemically reactive high temperature environment is conducive to copious production of free hydrogen when things get out of control, and recollect that the spent fuel pool is something not unlike a reactor core in its dotage. It is also likely that the containment structure itself can be vented to prevent excessive build-up of hydrogen inside. Both of these sources of hydrogen–the reactor core itself and the spent fuel storage facility–can and sometimes do result in explosive concentrations of the gas inside reactor buildings, and the multiple explosions that have occurred at Fukushima are almost certainly due to this source.
“Spent” fuel is actually a bit of a misnomer. Unlike ash from coal-fired plants, which is rich in radioactives and toxins like lead and arsenic, the “spent” fuel from a typical fission reactor still has quite a lot of energy generating capacity. When a nucleus splits apart during fission, it creates two or more “daughter” nuclei that are in general radioactive, beta-emitters with lifetimes of anything from nano-seconds to centuries. A significant amount of the heat generated by nuclear power comes from these decay products, and they don’t magically go away when you insert the control rods or extract fuel rods from the core. They continue to produce significant quantities of heat for days after a reactor is shut down, and after “spent” fuel is removed from a reactor it is typically left for months or years while these relatively short-lived fission products decay away.
This is why reactors have to be cooled even after they’ve been scrammed. The residual heat from fission products is more than enough to melt the core without ongoing cooling for several days to a week.
With regard to spent fuel storage, although beta-particles are not very penetrating, it is common for beta decay to feed a gamma cascade in the daughter nucleus, so there is quite a lot of gamma-radiation, which is penetrating, coming off the spent fuel as well.
What happened in Fukushima was a result of something that probably seemed like a good idea at the time: store the spent fuel adjacent to the reactor so the handling of the highly radioactive fuel rods could be minimized until they had time to cool down. This makes good sense until you realize it means putting the spent fuel into an elevated pool that could crack and leak in a large earthquake followed by a tsunami followed by one or more explosions from built-up hydrogen, leaving the “spent” fuel exposed and in the worst case capable of working up a little chain reaction on its own.
Finally, Japan is one of the few countries in the world that has invested in reprocessing. As I said above, the spent fuel has a lot of energy left in it. Some of that is in the form of plutonium isotopes that are themselves fissile, and can be chemically extracted from the spent fuel and reprocessed into new fuel rods. One of the Fukushima reactors was running “mixed oxide” fuel of this type.
Summary of Facts
So my take on the facts looks like this:
1) The Fukushima reactors are well-designed for their era, and have sufficient passive safety features that loss of core containment has never been a significant risk in the current accident.
2) It is very likely that four of the six reactors at Fukushima have sufficient plastic deformation of their cores due to heating from fission product decay that they will never be restarted.
3) Hydrogen production and build-up within the reactor building has resulted in multiple explosions and created significant risks to workers.
4) Release of radioactives from spent fuel storage facilities, probably due to fires ultimately started by exploding hydrogen, has been the primary mode of dispersal of radioactives into the environment, and while there is a minor public health hazard associated with it for people in the vicinity (20 – 30 km) there is no widespread risk and unlikely to be any. Some of the workers in the plant have almost certainly faced significant exposure, although at doses most likely to increase their statistical risk of cancer down the road. There is zero risk to anyone outside of Japan, and nearly zero risk to almost everyone living inside Japan. More people have been endangered world-wide by taking inappropriate amounts or forms of iodine than have been put at risk by the reactors themselves, even counting workers hurt in the explosions.
5) The spent fuel storage facilities at one of the reactors may have been at risk of re-criticality. This danger now seems past, and would only have been possible in the first place if there was either a major design error in the storage facilities or if the pool was significantly over-filled due to incompetence or malfeasance.
Where does this leave us? What is the “root cause” of these issues?
Any critique of a technology or social policy must ultimately come down to some claim about why the thing is bad, and then some claim about what is to be done. For example, socialism is bad because in the absence of market mechanisms of price-signalling socialist managers lack the information required to make rational decisions. Corporate capitalism is bad because the privileging of one form of social organization (the limited liability corporation) allows corporate managers to hide behind the legal skirts of the Nanny State while they screw the rest of us. And so on.
Critiques of this form allow one to formulate rational options to correct the deficiency (a social democratic mixed economy in the first case, a liberal democratic mixed economy in the second.) It is extremely rare to come across a critique that is so compelling that one can say simply, “Don’t ever do that.” This does happen, though: the death penalty is never justified as a social policy, for example.
When arguing policy we have to choose between viable futures, and people do not always agree on what is viable. I’ll get to this in more detail below, but first, what is my claim of the root cause of this kind of accident?
The energy density in reactor cores is extremely high. This is, after all, the whole point. Coal produces around 20 MJ/kg of thermal energy, so to generate 2 GWth–which is about what you need for an electrical plant with an output of 1 GWe–you need to burn 100 kg of coal per second, or 6000 kg (6 tonnes) per minute. The standard hopper car on a train holds 91 tonnes of coal (100 tons), so a 1 GWe plant is burning a rail car full of coal every fifteen minutes.
Reactor cores, on the other hand, are refueled every 12 – 18 months with typically 1/3 of the rods replaced each time, requiring a reactor shutdown of about a month. Thanks to their horizontal design CANDUs are notable for being able to refuel while running, as opposed to vertically oriented LWRs. However, CANDUs also suffer from failures due to sagging tubes as a consequence of their horizontal design…)
That means there is the equivalent of three to five years worth of coal-fired energy stored inside a nuclear reactor at any given time: that’s over one hundred thousand hopper cars of coal packed into a space a few meters on a side, waiting to be released as heat.
With that kind of energy density it’s no surprise that reactor cores melt a little now and then. And given that at least 30% of that energy is left in the fuel when it is removed (we don’t run to 100% burnup) the spent fuel storage facilities are a bit like huge piles of raw coal waiting to ignite. Not easy to get started, but no fun to put out should something happen.
Humans sometimes make mistakes. I know I do. Furthermore, given enough time, this tells us that “anything that can happen, will”. Any scenario with remotely non-zero probability will occur. Engineering failure analysis is a well-developed field, but it often neglects to take full account of human behaviour, simply because it is so difficult to think like someone else who is lazy, inattentive and rushed.
In the case of Fukushima, there have been at least suggestions in the news reports that the spent fuel facilities were over-filled, and the relative lack of containment of the spent fuel seems to me a very significant design error in the overall reactor ecosystem.
Those two factors: high energy density and human behaviour, seem to me at the root of all the nuclear accidents the world has experienced. I haven’t said “human error” because that makes it sound like something that could be changed. Humans have pretty well-known and well-understood tolerances, and while we can engineer out a great deal of susceptibility to certain kinds of behaviour, they aren’t ever going to go away entirely.
Nuclear power has the same problem Margaret Thatcher purportedly had with the murderous nutjobs in the IRA, who for some completely obscure reason thought that killing people might be the most efficient, effective way to bring about political change: they apparently said that Thatcher had to get lucky every time to stay alive, and they only had to get lucky once to kill her.
Because of the high energy density in reactor cores and spent fuel storage facilities, we only have to get unlucky once to create a hell of a mess. The mess is mostly economic in nature, although economics being what it is human lives always come into it in the end. If one of the Bruce or Pickering plants that power Ontario were to go down we would be facing the same kind of brownouts parts of Japan are now dealing with, and loss of power causes loss of life in a multitude of circumstances.
Anyone who believes that being against something constitutes a useful or interesting position on any social policy question is missing the point. Being against the death penalty is pointless if you aren’t also for something: rehabilitation, restitution, incarceration, whatever. Being anti-war or opposed to political violence is literally pointless if you can’t point to the many and varied and practical alternatives.
Thus, when considering the future of nuclear power, we can only do so in the context of practical alternatives, which to my mind break down as follows:
1) Use less power
2) Make nuclear more economically viable
3) Make other alternatives more economically viable
Using less power is problematic.
My carbon footprint is small for a Canadian. I walk to work and drive anywhere so rarely I burn a tank of gas every three months or so, I live in a small, well-insulated house heated with a high-efficiency gas furnace, don’t have an air-conditioner, and my primary forms of outdoor recreation are sailing, canoeing and hiking.
I still use a hell of a lot of power: somewhat under 500 kWhr/month, which is about half the average. But I really don’t want to make the lifestyle changes required to go much lower than this, thanks. I live the way I do because I like it, and that is the only way of living that is sustainable. People who believe that others should be forced to live some particular way do not believe in sustainability. Force is not sustainable.
With fancier power control and a few other fixes it might be possible to bring my energy consumption down by another factor of two without impacting my lifestyle. That would be great: if everyone did that Ontario would have no need to build new generating capacity for a couple of decades at least.
That’s fine in the context of Ontario, and if Ontario wanted to put in place a non-linear price regime that made the first 250 kWhr per month dead cheap and then applied a very steeply rising premium after that, or instituted “pay-as-you-go” billing, which was shown in the Guelph experiment to significantly reduce usage, I would be all for it.
But in the context of the world that isn’t fine. As a Canadian on the low end of the energy-consumption scale I am still using vastly more power to maintain my very pleasant lifestyle than the average person in the developing world. Simply telling people in China or India “You can’t have that” won’t do. It is not in the range of practical alternatives.
Making nuclear power more economically viable seems to me an option that is worth exploring. I am not particularly worried about nuclear safety, and anyone who is hasn’t been paying attention. The worst known nuclear disaster in the world–Chernobyl–has left the wildlife in the “dead zone” around the city much better off than in adjacent areas. When putting things in context, it is worth remembering that the worst nuclear accident in history has proven to be less damaging to wildlife than simply having humans around. The wild flora and fauna of the Chernobyl region are better off living in a “radioactive wasteland” than in a human-inhabited landscape.
The question is: can nuclear power be made economically viable? The answer is: probably. (Hopefully some corrupt Conservative–but I repeat myself–cabinet minister won’t come along and insert a ‘not’ into that sentence!)
There has been some research done on reactor designs over the past thirty years, and we have a number of ones that look promising, from helium-cooled pebble-bed reactors to fully-encapsulated designs that are never refueled, which have a lifespan of thirty or fifty years and are then disposed of, more like nuclear batteries than conventional power plants. The latter technology seems to me to have the fundamentally correct approach: engineer out all the operational stuff that is what results in the problems. If reactors like that had been in place in Japan they would have been geographically dispersed, so less subject to all getting hit at once, and the ones that were hit would have reacted like the big concrete blocks that they are. There might have been some radiation release into the environment, but it would have been trivial and nothing like the massive distraction from the really important relief efforts that Fukushima mess has been.
With regard to alternatives, I’m all for them, but doubtful of their economic viability, particularly with various anti-environmental groups opposing the development of wind and solar power because it doesn’t meet their unsustainable agenda of control. Solar has much greater viability than anything else, and algal biodiesel is my best bet for something to solve our liquid hydrocarbon problem in the wake of Peak Oil, but none of those are quite ready for prime time yet.
I do not consider coal a fuel worthy of consideration: it is dirty, dangerous and non-renewable.
Fear and panic are never helpful. Shouting “Fukushima! Chernobyl! Three Mile Island! Windscale!” does not constitute an argument, and the people who do it are the enemies of open, informed, fruitful discourse.
Nuclear power has a fundamental problem: the extremely high energy density in the core makes it susceptible to human behaviour, and gives it an unfortunate tendency to write itself off under circumstances where more forgiving energy sources would be subject to repair.
There are designs that deal with this issue pretty well, however, particularly sealed-unit reactors. I believe they have an important place in the future of the human energy generation landscape.
[Edit 2011-03-20 21:05 -- added some clarification around hydrogen gas production and corrosion in response to reader questions -- tjr]