I have mostly tried to avoid thinking much or having an opinion about nuclear power. This has been made easy by its minor and stagnating role in the global energy system.
But the idea of a nuclear renaissance is now competing for attention and funding in the low carbon energy debate and is becoming harder to ignore.
This post is an attempt at sorting my own thoughts on the topic in particular in anticipation of a heated political debate on he future role of nuclear power in Switzerland.
Not being an expert in nuclear technology, my opinions are based on the application of general engineering principles and best practices as well as some experience in building high-availability systems.
While nuclear reactors might be among the systems that are most obsessively engineered and operated for safety, nothing is ever 100% safe. Whoever promises otherwise is either lying or lacks imagination.
There is a general rule in engineering - that whatever can go wrong will eventually go wrong if you just try enough times - i.e. running something at large enough scale for a long enough time.
The question is what will be the consequences when things start to really go wrong? And how hard would it be to fix or revert them?
Worldwide, there have been about 600 currently active and retired nuclear power plants - 2 of which ended up in the center of a 30km nuclear exclusion zone for the foreseeable future. Which is about a chance of 1 in 300 or the lifetime odds of being shot to death in the US.
The picture above shows the potassium iodide tablets sent to us by the Swiss government because we live within 50km of one of the three nuclear power plants - along with 5 million other people, more than 50% of the Swiss population.
As an engineer, I would not want to be responsible for building or running a system where the damage from what can go wrong in the worst case would be bad enough to potentially bankrupt a small country or displace a population of millions for generations.
We should favor technologies which are resilient, fail safe and mostly harmless even for catastrophic failures. And no matter what happens, cleanup and dismantling should always be feasible with the same order of costs as building it in the first place.
There are inherent limits to the safety of a system which stores large amounts of energy in a small space, specially in cases when this energy is released uncontrollably. Yet we should still try to limit the amount of irreparable long-term damage caused by such incidents.
Admittedly the current reactor designs might be overly complex and not optimal from a safety-by-design perspective. They were to some extent a byproduct of the many billions of dollars in funding which the militaries of nuclear superpowers spent during the cold war in order to get the technology to process fuel for nuclear weapons or to power submarines and aircraft carriers.
One great hope of the nuclear industry are supposed to be Small Modular Reactors (SMR) - all kinds of novel designs, simpler, cheaper and mass producible in larger volumes. And which could eventually be improved iteratively by trial and error. This approach is supposed to replicate the success stories of how for example microprocessors have come to dominate nearly every field of computing - including supercomputers or how solar cells, wind-turbines and batteries are in the process of revolutionizing the energy sector or how a startup like SpaceX has become possibly the most competitive space flight operator. Startups often innovate by taking outszed if not reckless risks - not sure if this culture of disruptive innovation ("move fast and break things") is particularly suitable for building nuclear reactors, where breaking things can have long lasting consequences.
In some domains like computing, communication networks or renewable energy, building highly reliable systems by interconnecting many identical small cheap and typically less reliable modules has been greatly successful. But these architectures typically come with tolerating failures of components in the system. Not sure this is the right approach with a technology where each catastrophic failure could badly affect the surrounding areas and render them uninhabitable. Having many thousands of smaller reactors rather than a few hundreds of larger ones would mean a likely catastrophic failure every few years or maybe even more frequently as we might not be able to maintain the same rigorous operating and maintenance standards over a much larger number of sites as easily.
Maybe a more philosophical or ethical dilemma is the disposal of high-level, long duration radioactive waste. These need to be safely stored away for thousands of generations far beyond the cultural horizon of human history. Even if we end up building enough and suitable geological storage - how do we ensure that treasure hunters or archeologists in some far future thousands of years from now do not end up digging them up again? The timescales into the future these sites would need to be untouched would reach as far forward as has passed since the neolithic structures of Göbekli Tepe or Stonehenge were built - fortunately neither of these contained anything as dangerous, but we would not have remembered or understood what to do with it. How can we leave a responsibility like this for generations as far into the future?
The cold rational argument in favor of nuclear technology is that despite the 2 catastrophic incidents so far, relatively few people have died - relative to deaths for example from air pollution or climate change even up to today. Even if the cost is dozens of more nuclear no-go zones for generations, this might be a price worth paying compared to what we are currently doing to the planet in terms of burning fossil fuels to generate electricity. In particular countries with plenty of unpopulated areas might think differently about these concerns.
I admit that my reservations against nuclear power might be at least partially irrational. However they are also at least in part vindicated by the availability of credible alternatives. Alternatives like low-carbon renewable energy from wind, water, sun and storage, which might require more creative approaches on how to build and operate a future power grid or which might require us to accept changes in our habits and in our surroundings. We should give these alternatives a fair chance before we resort to solutions which can result in serious and hard to revert consequences.
Too much attention on next generation reactor designs that might or might no happen decades in the future is also drawing away focus, funding and attention from technologies which exist today and are ready to scale. While globally there are currently about 437 nuclear reactors with a total generating capacity of 368 GW (+3GW in 2023), the world has added 473 GW of renewable generation capacity in 2023 alone - most of it cheaper than any existing alternatives.
No matter my reservations against next-gen nuclear technology, I am most acutely worried that the promise of salvation decades into the future from a not yet existing solutions is going to be used once again as an excuse to avoid doing what can and should be done right now.