Small Modular Hallucinations

The self-refuting concept behind small modular nuclear reactors

Installing the external dome on the Linglong One “small modular reactor” in Changjiang, Hainan province, China. Source

Despite the narrative set out by nuclear advocacy groups, so called “small modular reactors” (SMRs in short) are neither small, nor modular — let alone affordable or scalable. Even as the idea of building such power plants is gaining prominence among business leaders and government officials alike, the economics of the concept almost never gets a mention beyond a few superficial statements. Recently I came across a terrific interview with physicist and author M. V. Ramana, who after demolishing a number of myths around nuclear energy, finally mentioned the bogus economics behind building and operating such reactors. As an engineer who spent two decades in and around manufacturing, supply chain management and product development I felt this is the time to put my two cents in, and elaborate further on the manufacturing and logistics aspects of the SMR-hype. So, here it goes.

Thank you for reading The Honest Sorcerer, and special thanks to those who already support my work; without you this site could not exist. If you are new here and would like to see more in depth analysis of our predicament, please subscribe for free, or perhaps consider a paid subscription. You can also support my work by virtually inviting me for a coffee, or sharing this article with a friend. Thank you in advance!

Small?

First, let’s start by clarifying what a small modular reactor is and what it’s not. For starters, it’s not small. At all. Based on their physical dimensions these beasts are almost as big as your regular run-of-the-mill nuclear power plant. While there are rumors about micro-reactor designs and military prototype reactors the size of a few shipping containers, you would not enjoy a long and happy life after spending more than a second or two in their vicinity. You see, there is no way of squeezing the necessary cooling and radioactive shielding (not to mention a containment vessel) into these dimensions... All of which is kinda necessary to keep the results of the nuclear reaction inside while transferring waste-heat into the outside, as opposed to contaminating an entire region Chernobyl-style. The only reason to call SMRs “small” is their electric power output, which is usually below 300 MW in contrast with regular nuclear power plants (NPPs) which are in the range of a 1000 MW — or 1 GW.

Illustration of a light water small modular nuclear reactor. Let’s face it: it is nothing but a gigantic water boiler with a steam turbine and a generator attached. (SMR) Source

In 2024 there were only two “small modular reactors” in operation, in addition to two high-temperature test reactors according to the IAEA, the International Atomic Energy Agency. The first one, the Akademik Lomonosov floating NPP, was built in the Russian Federation with two KLT-40S reactors of 35 MW(e) each, and was refueled for the first time in 2023. In operation since May 2020, it is supplying heat and power to the extremely remote town of Pevek in the Chukotka region; a port for the export of minerals as part of the expanding Northern Sea Route.

The other SMR operates in China as a demonstration unit at the Shidaowan site. It started commercial operation during December 2023 and generates 200 MW at full power from two reactors linked to a single power turbine. Other “small modular reactors” — under construction — can be found in Argentina (the CAREM-25 prototype reactor), in China (the ACP100 multi-purpose demonstration unit, also referred to as the Linglong One) and in the Russian Federation (the BREST-OD-300 in Seversk, Siberia). Fun fact, and as for a further illustration of the size of these beasts, the steel reactor base plate of the BREST-OD-300 reactor alone weighs a whopping165 tons; enough to build 143 cars from.

According to the IAEA there is only one modular reactor under construction in the U.S., the Hermes Low-Power Demonstration Reactor in Oak Ridge, Tennessee. It is the first and only Generation IV reactor to receive a construction license from the Nuclear Regulatory Commission, and the first non-lightwater reactor permitted in the USA in over 50 years. However, it will not generate electricity. Instead, this fluoride salt-cooled, high-temperature reactor is being built to demonstrate affordable heat production. (With that said, it won’t be able to generate high enough temperatures to melt steel, glass or even copper, let alone to make cement.)

Modular?

The key concept behind SMRs is their much touted modularity, “the quality of consisting of separate parts that, when combined, form a complete whole”. Well, based on this simple categorization every nuclear reactor is modular, coming with many interchangeable parts from valves to pumps, back-up generators and the like. None of the NPPs operating around the world were shipped as a whole unit, but assembled on site — simply due to their massive size. As we have seen above, so called “small modular reactors” are no different in this regard, with Admiral Lomonosov coming closest to being “shipped” in one piece — pun intended. So, what’s the difference then?

The alleged modularity of SMRs comes from the standardization of their major components, such as the reactor pressure vessel among other items. Then an industrial model of mass production — with a large proportion of factory prefabrication — can be applied to their manufacturing and voila: cheap nuclear is made available to everyone. Since building custom nuclear reactors on site is expensive, the concept does seem to make sense, but only on paper. Allow me to draw an analogy between vehicle component manufacturing and design (the area I’m most familiar with) and nuclear reactors here. What makes the manufacturing of cars, consisting of 30,000 different components cheap is not as much the uniformity of parts across generations or types, but the sheer volume of production. We are talking about at least several hundred thousand if not a million copies of the same part delivered to car assembly plants worldwide, where they get built into the vehicle you just ordered online.

Compare that number to the number of reactors ever built globally: 815. Yes, a little more than 800, with 440 of them still in operational condition and a further 100 planned. Not exactly thousands, let alone millions. Even if each and every one of these reactors had been built from exactly the same set of components, that would have meant less than a thousand reactor pressure vessels, a several thousand circulation pumps, heat exchangers and steam generators. Folks, these are nowhere near series quantities when it comes to manufacturing. Making parts (especially such high tech, high spec parts) is costly, but setting up even a semi-automated manufacturing line costs even more. Building up a specialized line for producing many hundred thousand copies of the same brake-booster pump built into cars is one thing. And while such a line might cost tens of millions of dollars, huge order quantities reduce the per part investment cost to a mere ten to twenty dollars, sometimes even less. Unfortunately, the same could not be told about high pressure reactor vessels made in the tune of hundreds at best, with much higher specs and a much-much bigger size. I’m sorry to disappoint you, but there is no way you can make reactor components cheaper by applying “mass production” techniques; the only way to produce them is to cut each sheet of metal individually, then weld them together either by hand or at best using a programmable welding robot (for simpler tasks). Just take a look at this video:

This takes us to the question who will build and operate these ostensibly “small” and apparently “modular” reactors then? Welding pressure vessels by hand, or programming KUKA robots takes a huge amount of skill and years of training. Producing pressure vessels and other components with a somewhat smaller capacity than a regular nuclear power plant uses does not necessarily reduce the amount of work needed to complete them. Sure, less material is used, but the amount of welding and assembly labor does not necessarily shrink in a proportional way.

Scalable?

The idea of SMRs defeats economies of scale on every level imaginable. As we have seen above, building reactors with a smaller nominal power output does not make them more modular than any other run-of-the-mill reactor. If these beasts are not made by the tens if not hundreds of thousands, little economies of scale can be reached by standardizing components other than screws, pipes and bolts. Simply put the volumes are not there to achieve cost reduction by automation, leaving us with our current, labor and skill intensive manufacturing techniques.

Another major cost element in building a nuclear power plant is the site itself. Finding the right site and obtaining permits for building a power plant there is a daunting task in and of itself. Doing the same thing for several smaller sites does not make things easier, to say the least, nor does it make things cheaper. That, however, is the lesser problem. The bigger issue is related to the supporting infrastructure: human resources, fuel management, waste disposal, buildings, equipment and transportation systems among many other things. NPPs, are no small task to operate.

Preexisting infrastructure is the reason behind having multiple reactors on one site (sometimes as much as six). Maintenance crews, engineers and managers can easily go from one reactor to the other to support activities like refueling, repairs, maintenance and so on. Spare parts can be stored in a central building and their stock levels can also be managed centrally. This naturally generates efficiencies of scale: you don’t need six managers, six maintenance crews or six spare parts for each reactor, as you can easily move things and people around. Imagine doing the same thing for six SMRs, scattered around the country with a hundred or more miles between each location. Planting an SMR here and another there is not how you save on personnel or maintenance costs. Quite to the contrary.

This takes us to the final question: where the engineers and other highly skilled people are going to come from? Welding pressure vessels, operating reactors or maintaining them is not something you learn on an online course in a day. It takes years of training and practice — and as long as it will be easier to find a well-paying job as a marketing manager young people will not take the extra effort to pass tough math and physics exams. Most worryingly, though, it’s not only the will which is lacking but skills, too. STEM skills are on the decline for years now:

National assessments show a sharp decline in elementary and secondary student mathematics performance since the COVID-19 pandemic. From 2019 to 2022, average mathematics scores of fourth and eighth grade students dropped to levels last measured approximately 20 years ago.

The situation is not a tad bit better with literacy rates either: one in five U.S. adults (21%) has difficulty comparing and contrasting information, paraphrasing, or making low-level inferences and thus considered illiterate. At such high rates of illiteracy what can we expect from the rest of the population…? I’m sure the last thing you want to see is people scratching their head over an emergency shut-down protocol while their colleagues scramble to bypass an AI customer support agent on the other end of the line:

“I can certainly check that for you. May I put you on a brief hold?”

In the meantime the literacy rate of China has reached 99.83% in 2021. Contrary to common wisdom, employers no longer go to China for cheap labor, but for skills. Be it jobs requiring high manual dexterity or higher education China can offer both — no wonder they are zooming ahead in all technological areas, including nuclear. Chinese universities train engineers by the millions these days — 1.7 million in 2023 alone — whereas the U.S. produces less than a tenth of that figure: 120,000 STEM graduates in a year. Food for thought. Still, despite all that, not even China plans to roll-out nuclear en masse. By 2050, China targets nuclear energy to account for 15% of its electricity generation, which is a lot more than many other countries, but still much less than what is going to be generated by wind and solar by then.

Last but not least, and contrary to what nuclear advocates tell the public, SMR-s are less safe and produce more waste than regular NPPs. To cite a study published in PNAS on the topic:

The low-, intermediate-, and high-level waste stream characterization presented here reveals that SMRs will produce more voluminous and chemically/physically reactive waste than LWRs, which will impact options for the management and disposal of this waste. Results reveal that water-, molten salt–, and sodium-cooled SMR designs will increase the volume of nuclear waste in need of management and disposal by factors of 2 to 30.

How is that scalable?

Photo by Dan Meyers on Unsplash

Conclusion

SMRs are a perfect example for “the wrong answer to the wrong question.” They are neither small, nor modular or cheap, let alone sustainable. Most importantly, however, they miss a crucial point. You see, the predicament our civilization faces — besides an unfolding ecological collapse — is a peak and decline in the availability of affordable resources, most notably crude oil. This is an issue nuclear power plants offer no viable solution to, as producing more electricity provides no answer to the loss of natural habitats or the depletion of resources, including uranium itself (1). (Besides peak farmland we might have already passed peak uranium, too, just sayin’.) Not to mention the fact, that neither mining nor agriculture or long distance transport could be technically electrified at the scale and at a speed needed to avoid a major economic dislocation from a peak and decline in fuel production. And without adequate supply of diesel fuel, how on Earth do we plan to continue mining and delivering the uranium, copper and all the gazillion other materials needed to run a nuclear power plant? And if these technologies cannot be maintained properly then pray tell, why do we need more nuclear power plants?

The issue our high-tech civilization should be grappling with today is not “How do we generate more electricity?” but how do we reduce the overuse of resources and the pollution which comes with that — aka overshoot. Sure, using electricity from nuclear to turn the planet’s dwindling natural and mineral resources into dead matter destined for the landfill is certainly more climate friendly than burning coal. However, nuclear won’t solve neither the ongoing ecocide, nor pollution in general — let alone the slowly unfolding affordability crisis of heavy transport fuel, making all the mining, agriculture and construction — and destruction — possible. The long decline is upon us, and no amount of hopium in the form of “small modular reactors” will be able to halt it on its tracks.

Until next time,

B

The Honest Sorcerer is a reader-supported publication. Please consider a subscription or perhaps buying a virtual coffee… Thanks in advance!

Notes:

(1) So called “breeding” reactors running on Thorium are still in development phase and nowhere near commercialization. The molten salts involved in the process are highly corrosive and demands custom-built alloys capable of withstanding both radiation and chemical degradation. Besides, the reactor needs an initial load of uranium-235 or plutonium-239 to start the chain reaction until enough uranium-233 is bred from the thorium. So while it seems technically feasible to build Thorium reactors, they will not save us from nuclear waste, nor can be scaled up to replace fossil fuels in all areas. We don’t have that much time left. Oil is about to start its long decline this decade — first slowly and tentatively, then ever faster — causing all sorts of problems preventing us from investing in these still to be matured technologies.