Does Australia need nuclear to meet our Paris climate targets?

Australia2040.com lays out a plan for securing and improving Australia's economy, energy supply and culture over the next twenty three years. The problems outlined in the website (e.g. overreliance on mineral extraction for economic growth, unsuitable education system for tomorrow's economy, unstable energy supply, health impacts from coal) are indisputable. Australia will have to change radically over the next twenty years if we want to enjoy a better quality of life than we have today. Many of the solutions outlined in the plan make sense to me. We absolutely do need to diversify our economy away from resource extraction and real estate appreciation. We do need to train our kids in the skills they'll need for a digital economy. We do need to replace our ageing coal power stations that are belching GHG emissions and causing health problems.

But..I find it hard to stomach the proposed solution for our energy crisis. The authors of Australia2040 advocate the construction of GenIII/GenIV nuclear reactors to provide stable baseload power. Their argument is that intermittant generation sources like solar/wind cannot replace coal because they are non-dispatchable. I decided to set aside my nuclear skepticism and explore those assumptions to investigate whether we do indeed need nuclear.

It comes down to seven questions:
1. Is nuclear safe enough to power the world?
2. Is nuclear significantly cheaper than solar + storage?
3. Do we have enough materials to build all the solar panels/batteries/nuclear power stations we'd need considering limited supplies of Lithium/Cobalt/Uranium?
4. Do we have enough physical space to build that much generation capacity considering NIMBYism?
5. Can the new generation be brought online quickly enough to meet our emission abatement targets considering typical build times for nuclear?
6. Will we end up with stranded assets if we build loads of solar/nuclear considering cost reduction curves?
7. Are there other benefits to nuclear/solar besides providing power?

Safety
Australia2040 argues that nuclear is safer than coal in terms of deaths/disability adjusted life years (DALYs) and is actually comparable to solar/wind. We must bear in mind that solar panels and wind turbines can catch on fire.

These incidents don't capture as much press as a reactor meltdown but they're nonetheless a hazard. I'm not going to quibble with Australia2040's argument that nuclear's safety record is comparable with solar/wind in terms of deaths per kWh especially when you exclude reactors that were known to be unsafe before the meltdown (e.g. Chernobyl).

However, deaths/kWh is not the only way to look at safety. When you consider the millions of liters of water that were contaminated by the Fukushima meltdown, the possibility of intergenerational health impacts and the potential for a malevolent actor to gain access to radioactive material (check out the numerous unapproved incursions Plowshares members have made into nuclear reactors), the safety stakes are securely positioned on the side of renewables. Even if GenIII/GenIV reactors do turn out to be much safer than previous designs, the potential for massive damage still exists. All it takes is for one once-in-a-century earthquake/a terrorist flying a plane into a reactor/an enraged dictator with an ICBM and your whole power system and the surrounding area is impacted for decades to come.

Score so far: Renewables 1, Nuclear 0.

Cost effectiveness
Currently solar + storage is significantly more expensive than nuclear. If we are uber pessimistic and assume we'd need 36 hours of storage for every kW of generation capacity, then renewables are an order of magnitude more expensive than nuclear because batteries/pumped hydro storage is at least $100/MWh right now. Storage would need to be 2-3 times cheaper to be competitive with the LCOE (levelised cost of electricity) figures for nuclear.

However, it's worth looking at cost curves. Here's the cost curve for solar:

And here's the cost curve for nuclear:

In the last forty years, the cost to build a kW of nuclear power has substantially increased whereas solar/battery storage are dropping by 21% per year.

This means within 10-20 years, we're likely to see price parity for nuclear vs solar + storage. Given that the South Australian + federal governments have already earmarked hundreds of millions of dollars for Big Storage, the price is already right for solar + storage and it's going to keep getting better. Here's a Google spreadsheet analysis of how nuclear will likely compare to solar + solar by 2030.

Another factor that comes in is the cost of natural disaster insurance for nuclear. With a predicted cleanup bill of up to $600 billion for the Fukushima disaster, insurers are not likely to let nuclear power station operators off easy even if they do have better technology. In some countries, nuclear power station operators are given liability caps and do not have to attain 100% insurance coverage for disasters. If they did, anti-nuclear advocates argue that nuclear would not be cost competitive.

Score: Nuclear 1, Solar 0.5 (because solar + storage has better long term prospects)

Tally: Nuclear 1, Solar: 1.5

Material availability

If we were to try to hit 100% nuclear/solar+storage with today's technology, we would quickly run out of requisite materials before we got close to the 100% mark. Uranium would run out if we generated power exclusively from current generation nuclear plants. We have 230 years of uranium supplies for the current generation fleet. But nuclear is only 1% of total energy requirements. If we were to ramp up to 50% nuclear, we’d only have ~4.6 years worth of uranium. Fast breeder reactors are 100x more efficient though so if we used them, we’d have several hundred years worth but this is more of a hypothetical currently as there are only two commercial scale fast breeder reactors in the world (the rest are experimental/already decommissioned).

Similarly, we would run out of Lithium, Cobalt, Tellurium and several other key ingredients before getting close to 100% penetration of solar + storage (link).

For either technology, we are going to need a step change in resource extraction technology (e.g. seawater extraction/better recycling) and generation technology (e.g. proven GenIII/GenIV reactors, saltwater flow battery technology, perovskite panels).

Score: Nuclear 0, Solar 0

Tally: Nuclear 1, Solar 1.5

Do we have enough physical space to build all the solar + batteries/enough non contested space for nuclear?

Nuclear is far more "geographically efficient" than solar/wind. Energy per sq km for nuclear is several orders of magnitude higher than for solar. The issue for nuclear is not the square footage but the NIMBYism (not in my backyard) that pervades most Australian cities. To deal with community opposition, you'd probably have to chuck a reactor far away from population centres. The Australia Institute came up with a plan in 2007 (prior to Fukushima) for where to place nuclear power stations including near Noosa, Port Stephens and Jervis Bay. All I can say is "good luck with that".

The problem with solar + storage is surface area. According to my analysis, that's not really a problem. We'd need 1% of the world's surface area for the solar panels and a negligible portion for batteries.

Surface area analysis:

1. 2020 annual energy demand = 184,297 Twh (https://www.eia.gov/outlooks/ieo/world.cfm) - including hydrocarbons for transport - we’re going to assume that either all transport gets electrified or we use a power to gas technology
2. = 21.03 TW peak demand (184297 / 365 / 24)
3. Assuming crappy conditions, assume need 36hrs worth of storage
4. Therefore world needs 757.08 tWh of storage and 21.03tW of installed solar
5. Surface area required for solar:
6. 1. .24 Sq m per 1kW (1kW per square meter of insolation and efficiency of 24%)
   2. Therefore need 5,047,200,000 Sq m (21030000000 kW * .24 sq m)
   3. 5,047,200 square km
   4. Total area of globe = 510 million km
   5. :. use 1% of globe for solar generation
7. Surface area required for batteries:
1. One power wall 2 is 6.9sqm for 200kWh
2. 757080000000 kWh / 200 kWh * 6.9 sqm
3. :. need 26 119 260 000 m2
4. :. need 26,119.26 square km

Essentially, surface area isn't really a problem for solar + storage. You might cop a bit of community opposition to wind turbines but PV solar is pretty unobjectionable.

Score: Nuclear 0, Solar 1

Tally: Nuclear 1, Solar 2.5

Can we bring new generation online quickly enough to reach GHG reduction targets?

Australia2040 argues that a new GW scale nuclear power station can be built in 5 years based on some recent projects in the UAE. However, the majority (35/55) of stations currently under construction are significantly behind schedule (up to 40 years for construction) and way over budget (link) due to technological and political issues. Furthermore, given that the UAE ground breaking ceremony took place in March 2011 and only one of the reactors is likely to be finished in 2017, Australia2040's figure of 5 years seems misleading. At best, you're looking at six years (and the first reactor hasn't been finished yet) and that's assuming that KEPCO (the firm responsible for construction) could handle multiple reactor buildouts at once. Given the shortage of trained nuclear construction engineers globally let alone in Australia, that's a big assumption. Realistically, once you factor in government approvals (it took 2 years in UAE and would presumably take much longer in Australia given community opposition), it's highly unlikely that a GW scale reactor would be built in Australia before 2030.

By that time, solar + storage will have improved further in efficiency and dropped further in price. With Elon Musk proclaiming that Tesla could install 300 MW of batteries within 100 days, I think it's safe to say that solar + storage annihilates nuclear in construction speed.

Score: Nuclear 0, Solar 1

Tally: Nuclear 1, Solar 3.5

Will we end up with stranded assets?

One of the biggest blockers to new coal stations is the concern that they will become 'stranded assets', unusable because it is more expensive to continue to run them than to build new gas/renewable plants. This is a issue confronting nuclear power stations as well. The US has a little under 100 operational nuclear power stations, many of which were built decades ago. A significant number are under threat of early retirement because they can no longer compete with cheap renewables/gas. The problem is that it's pretty difficult to build a plant that lasts 40 years and remains economically competitive the whole time. Technology is going to keep improving and before long, you'll get left behind. Even five years ago, solar was not cost effective without government subsidies. Now that panels are retailing for below $1/watt, they can hold their own, paying for themselves within a few years. The same phenomenon will happen with batteries as well.

Given the rate of technology improvement we've experienced in the last decade and the predicted advances that are due to come, does it really make sense to invest on a 40 year time horizon? Chances are that within twenty years, solar + storage will be so cheap that any nuclear reactors that have been built will have to be decommissioned at great expense.

Score: Nuclear 0, Solar 1

Tally: Nuclear 1, Solar 4.5

What about other benefits?

Nuclear reactors aren't going away any time soon. We need them for medical purposes, for sterilising potentially dodgey products before they come through customs and for powering rovers on far flung planets that need a lot of power in a small payload. Long term, nuclear fusion is likely to be what fuels humanity's expansion into the stars. Therefore research into nuclear reactor technology must continue. We also need to build fast breeder reactors to deal with nuclear waste from old school reactors. Other geological storage solutions are unsafe and expensive.

On the other hand, solar + storage has an inherent advantage in its geographical redundancy. With nuclear, if one power station goes down, the nearby area and possibly the national grid is going to be in danger of blackouts (a la Japan post Fukushima). With solar + behind the meter storage, every household has a buffer against outages and can deal with an extended period of grid unavailability without any consequences. Microgrids are the future of electricity distribution.

Score: Nuclear 1, Solar 1

Tally: Nuclear 2, Solar 5.5

Conclusion

Neither nuclear nor renewables are suitable for decarbonising the world's energy system with today's materials. We will need technological innovation to make that happen. Currently, nuclear has advantages from an economic perspective over solar + storage but within twenty years, renewables will be so much safer, better, cheaper and more reliable that they will outcompete nuclear. Given the massive hazard posed by a reactor meltdown and the comparatively slim benefits (cheaper baseload power), nuclear represents an asymmetric risk that should not be taken in Australia. 

Nuclear does have its place: other countries that have built nuclear reactors in the past will need fast breeder reactors to deal with their transuranic waste products. Australia is not one of those countries. The only scenario which will warrant nuclear power in Australia is if energy storage has not dropped in price by 200% by 2025. If technology improvement slows, then Australia's only option to meet emission reduction targets will be to build nuclear reactors. At present, this contingency does not seem likely. Solar + batteries are nicely tracking a 20% price reduction for every doubling of production learning curve.