Civilization is a long game
There are only two real currencies in the world: energy and material. Conservation efforts are laudable, but it is literally impossible to conserve to zero. The old ways of collecting energy in the form of carboniferous fuels has proven unsustainable – if not for finite supply, for climatological side effects. If we can’t conserve to zero, we have to work out a way to replace our use of carboniferous fuels.
Renewables are a nice thought here – the disc of the earth receives on the order of 175 petawatts of power from the sun on a reasonably constant basis. This largely goes to maintaining earth’s thermal equilibrium – about 122.5 PW are absorbed into the atmosphere and ground, which re-radiates about 122 PW back. (Incidentally, there’s a slight energy excess that’s been growing – about 500 TW – which is the drive of our current climate change. Lower our atmospheric carbon, and the atmosphere insulates less well, allowing us to re-radiate that half a petawatt). This energy not only warms us all, but drives wind currents, evaporates rain to be redeposited to higher elevations, creates waves and wind currents – all renewable energy is ultimately powered by the sun.
This seems like it’d be a nice, consistent flow of energy we could use – except it isn’t. The rotation of the earth, the irregular terrain, and the sheer scale of the system compared to our scales, means that the fluctuations you’d expect to be small turn out to be huge.
This is a problem: mankind’s energy use variation is not so diurnal, nor so chaotic. We require a baseline level of energy to run our streetlights, our trains, our factories, our security systems, our computers, fridges and freezers, climate control systems, and a thousand other things.
We never fully switch off – but the tremendous, diffuse energy that the sun provides us does, and it does for hours at a time. Worse, all these systems that the sun influences interact in interesting ways – so on occasion you’ll get an entire windless, sunless week in any given location. Moreover, during different times of year, you’ll get a fraction of your average output on a long timeline, resulting in the same effect as a long stretch of dark, windless days.
Even assuming zero transmission losses and complete international connectivity, we’d need about a week’s storage at the 16 TW level the world uses in electricity – about 2.6 petawatt hours.
If we’re to rely solely on diffuse, intermittent energy, we would need to undertake an energy storage and connectivity project that would make the US highway system seem like a toddler with a new train set.
Assuming that project is feasible and successful, our best chance for storage is to use lead acid batteries – summing up the total available lead resources available to mankind leaves its potential capacity at the top of the list (Lithium, while higher capacity per atom, is sufficiently rare that lead beats it out by a couple orders of magnitude).
Using the entire planet’s lead resources to store energy for the grid would suffice for about 6.5 hours – we need a week, and we need it worldwide.
The nuclear problem
So we need a stable, baseload electricity generator that produces as little carbon as possible. Fission is, at this point in history, the only technology that does this.
However, nuclear has a number of challenges.
Ionizing radiation, of the sort that is produced by the products of the fission reaction, is deadly to humans at high doses, and carcinogenic at even moderate doses. While it is true that the reactor itself releases quite a bit of radiation, which we capture and convert to heat and, ultimately, electricity, the radiation does not simply stop when the reactor does; it is emitted independently by these fission products. When they are released, any significant concentration of them poses a morbidity risk to anyone who is sufficiently – or worse, internally exposed.
The nuclear industry does an excellent job of containing radioactive material and, in cases where a failure to do so has occurred, keeping the public away and de-concentrating radionucleides until their presence is no longer a hazard. Indeed, modern commercial designs have elaborate, layered containment to prevent release of these radioactive materials at all.
And it is for good reason that these are elaborate: the matter that needs to be contained is at extraordinarily high pressure and moderately high temperature. The physics of a reactor core is that it very much does not want to be where it is.
The result of this is that each nuclear power plant built using such a technology is necessarily a one-off design: siting very much dictates where and how a plant can be situated, so as to minimize the possibility of radioactive release and, failing that, to guide any releases away from population centers and get it diffused thinly into the environment as quickly as possible. The plant, buildings, and core all have to fit tightly together and with the terrain for maximum safety in the event of disaster.
This means that nuclear plants take, on average, from eight to ten years to build, utilizing a team of people specially trained with working on all sorts of high-test materials. There are only enough people with the right skills to do this to work on about 60 reactors annually, giving us the ability to create only 6 GW of new capacity a year.
We have roughly 14 TW of fossil-driven electricity to replace – to get it done in 30 years, we need to complete construction of a little over a gigawatt per day. To this end, we need standardization and assembly-lining of the construction of nuclear power plants. We need to be able to build them with common, skilled labor in the space of a year a piece.
The problem with renewables is that their energy is not readily dispatchable; the problem with pressurized nuclear is that the generators are not readily deployable. For a technology that we need this badly, it would be impossible to roll out conventional nuclear in a time-frame that would do anything useful in the face of climate change: we just don’t have the trained manpower.
One possible solution to this that is being explored is to build reactors that do not require pressurization. The main driver for having pressure in your reactor is not the nuclear energy at all, but the water we use to move that energy around. For the plant to get anything like a reasonable efficiency – and therefore, power output – the water has to be super-heated well beyond its atmospheric boiling point – meaning that you need a very strong container for the core.
If the core is breached, not only is the pressure released, and the nuclear rods left uncooled and melting down, but the core’s inventory of water flashes, very suddenly, to high pressure steam. This necessitates a large, strong containment dome around the core to keep all that nasty stuff in while it’s superheated and trying frantically to get out.
Both of these macroscopic parts require special training to build, weld, and fit, and there are a very few people in the world that can do these things.
By replacing the coolant with something that has a much higher boiling point – proposals so far have been lead, sodium, and various mixtures of salts – we can enable a low pressure system, one that any competent dock machinist can fabricate parts for (they build things of similar scale, complexity, and quality standards every day in shipyards). The primary containment can be weaker and tighter against the core and have the same safety margins – reducing costs, expertise, and specialized equipment. Further, since the liquid phase of the coolant is so far outside the range of normal atmospheric temperatures, we avoid the rapid phase change that necessitated that large dome – it can be a reasonably small concrete bunker.
Meanwhile, when you’re not working with pressure, the specifications for virtually every penetration into and out of the “nuclear island” – e.g., the primary and secondary containment – drop dramatically. You don’t need specialized valves that can move against 800 psi. You don’t need blow-off valves in case your specialized valves stick. What you do need is corrosion and heat resistance – but these are things we know how to do from the fossil fuels and chemical industries.
Further, as we eliminate the strongly macroscopic parts in favor of smaller, shippable, machinable parts, we enable the assembly line production and standardization that we desperately need if we are to become dependent on nuclear.
And we should. For the reasons stated above, we need nuclear, badly – but more importantly, we need a new nuclear. However (and I disagree with many on this subject), while we need a new nuclear, we don’t need new technology. We need to implement existing technology. We can leverage knowledge we already have to attain our mid-term goals, and avoid “boiling the ocean” with inviting, but ambitious new technologies.
Taking the long road
People talk about fusion and breeders and thorium – and these are awesome and promising, and I hope to see them in deployment in my lifetime – but we don’t need a new fuel cycle (yet). We have enough known uranium reserves to last us about 175 years at 14 TW, growing at a rate of 5 TW per decade, even if you only count the tiny fraction of it that’s fissile. Fusion and breeders and thorium can be an economic advantage for another day.
We need to leverage one of the non-water coolants we have some experience with, e.g., a molten metal or salt. We need to do this so that we can build reactors at a rate that would do us some good in the fight against climate change.
And the economic argument is sound: any certified reactor running at atmospheric pressure is going to win the day on costs and safety. Commercialization of such a device will be the first of a few oncoming events that will wholly reshape the nuclear industry. The first to market with such a device will quickly dominate, and their first product will make up a large proportion of the replaced fossil electricity capacity, quickly overtaking the regulatory costs of designing a new reactor.
Another somewhat less-tested technology that we can subsequently leverage to further reduce costs is to dissolve the fuel in the primary coolant, making a homogeneous, fluid-fueled reactor.
A solid fuel has a very high temperature gradient, which limits its power density; a dissolved fuel doesn’t, meaning you can make the core smaller. Further, a dissolved fuel will readily give up the types of nuclear ash that most suffocate the reaction – namely xenon and other noble gas isotopes – thereby lifting the limits on the amount of the fuel that you can actually burn. These limits are what drive the size of the waste stream for a conventional reactor, and lifting them reduces a reactor’s waste profile significantly.
Moreover, a high-temperature liquid coolant necessarily has a high freezing point, enabling any reactor that dissolves their fuel into their primary coolant the ability to defuel their core in seconds, automatically, in the result of any failure using another technology that’s easily understood and has been tested in a nuclear environment: the “freeze plug”. Basically, you use your plant’s electricity to drive a fan to keep the drain out of the bottom of the core cold enough to freeze. Core overheats? plug melts, core drains. Loss of coolant? Loss of fuel. Catch it in the drain tank. Loss of electrical power? Plug melts, core drains.
The first fluid-fueled reactor is what’s going to be the second game-changer for the nuclear industry, on grounds of pure political hay: walkaway safety and a seriously reduced waste profile. By the time a fluid-fueled reactor is commercialized, you’re going to have radiophobes everywhere concerned about the impact of their newly built local atmospheric-pressure nuclear plant (because, at this point, there should be a LOT of them). In order to undercut the political fights they’re now waging just to do business, utility companies will flock to the new, safer, cleaner alternative.
I haven’t mentioned fuel costs, and with good reason. The energy density of fission means that fuel costs are easily the smallest influence on the equation. It’s one of the reasons that conventional reactors can survive with well under 10% burnup and expensive fuel assemblies: they can simply waste 90% of their fuel without appreciably affecting their bottom line.
It’s at this point that thorium and breeders will begin to become important, for similarly political reasons. Questions will arise (far more ubiquitously than they are now) about what to do with the old stocks of spent fuel, and about the sustainability of uranium. At this point, with sodium and / or salt technology now well understood in deployment, the problem set facing thorium and fast breeders has been largely solved.
Even if it takes 100 years to get to this stage, we’re well within our sustainability limits, and have the time available to move on to more ubiquitous fuel cycles. I hope it doesn’t take this long, but should it, we’re still in the green.
I’m confident that fusion will come along at some point in the billion-or-so-year time-line before thorium and uranium become scarce assuming breeder technology.
I hope this article has some influence on the nuclear industry’s R&D efforts. If I were to distill my essential message to them, it would be this:
Moon-shot towards an atmospheric pressure reactor. Any basis; it doesn’t matter.
You can fight over those innovations later.
Pressure is what is in your way.
Pressure is the worst point of failure for a conventional reactor.
Pressure is what drives your costs.
Pressure is what the public actually fears.
Collaborate if you have to.
Beg for research grants and agree to open-source the technology if you have to.
Get atmospheric pressure first, and save the rest for later. That’s what will enable fast builds, safe reactors, and a thriving nuclear industry.