Max Power 30 July 2023

Cameron,
I could be wrong, but I think that you may have misread the Statista figures for PV manufacturing.
My reading is that the current figure is a bit under 250 GW, not MW...

Cameron MacPherson 27 July 2023

While I don't disagree there are several obstacles to making SMRs commercially viable, I found this article had several omissions that were curious coming from a former NRC chair. I also found it rather dismissive to reduce a tremendous volume of work from talented professionals as "Tech/Nuke Bro Libertarianism," as if the litany of SMR designs that have made it thus far were merely manifestations of overconfident, Patagonia-sporting young men searching for existentialism through a smoky haze of designer weed. Both of us know that assessment is neither true nor fair.

First, I'll touch on what the article omitted, the most significant one being that SMRs have been operationally viable for decades and during that time have provided the backbone of NATO's power projection and strategic nuclear deterrence. Every nuclear submarine fielded by the U.S. Navy is powered by a Small Modular Reactor under the U235 fuel cycle, the same is true with all 11 Nimitz-class carriers and every Ford-class carrier hereafter.

The Virginia-class submarine (SSN-774) is powered by a 210 MW S9G nuclear reactor and has a complete unit cost of $2.8 billion per-submarine. That's an overnight cost of $13 million per megawatt that comes stock with a *world-class attack submarine.* The Navy has been fielding SMRs for 50+ years without accident, so we know the technology is viable. Does that mean a startup can replicate their (highly classified) success? Not necessarily, to be fair. But we know with certainty the reactor designs work as intended, as the feat has been replicated more than 100 times. The technology in and of itself is categorically viable, highly reliable and demonstrably safe.

Whether the U235 fuel cycle, or thorium-fueled LFTRs (which, yes, are viable), or sodium-cooled, or fast neutron, we know that the point A to B is accomplishable. I would also speak, without getting too technical, that certain agency dismissals of alternate fuel cycles (such as Thorium-232 by the UK's NNL) that dispute its inability to make weapons apply their rationale in a vacuum, specifically, assigning proliferation risk based on potential fissile cross section (U233/NP-237) while ignoring that a sub-viable critical mass of either isotope would kill anyone attempting to work with it in less than two hours (whereas 80%+ U235 and P239 are safe to handle with minimal PPE). Practicality matters, ultimately - rogue states aren't going to procure the infrastructure necessary to avoid these problems, especially since neither of these exotic isotopes has made a militarily effective device (all the more so since their radiation would render inert the fast relay switches necessary for implosion, as well as any electronic PALs on the devices themselves).

I also must speak to renewables. As a fan of both renewables and nuclear, the schism between acolytes of each respective technology is frustrating as both stand to play important roles in a clean energy future. Yet the notion that renewables alone are up to the task is a challenging purchase.
Even if energy demand stayed constant - no electric vehicles - the world consumes ~25,000 terawatt hours every year, or 25 trillion kilowatt hours. This figure, broken down on a per-day basis, arrives at 68.5 billion kWh that needs to be generated per-day to meet that aggregate total.

Ignoring storage, and assuming a global average of 5 peak sun hours per day, a single 400-watt panel would output 2,000 watt-hours per day, or 2 kilowatt-hours. That's 34.25 billion 400-watt solar panels that would need to be manufactured, deployed and wired in just to meet *current* demand. If electricity demands increase by 30, 40, 50% as expected due to increased EV's and AC use due to climate change, the number of solar panels increases proportionally.

While I am again a big fan of solar, and believe solar can actually play a much larger role than it does currently, the idea that humanity is going to build 35+ billion solar panels (at minimum) with a lifetime of 30-40 years at a manufacturing material throughput of 17,000 metric tons per TWh and wire it all in with an unquantifiable volume of copper is difficult to take seriously. Indeed, according to Statista, the world produces ~250MW of solar panels every year, which at 5 peak sun hours carries an output of 1,250 MWh a day. To meet the target set above, we need to generate 68.5 million megawatt hours per day, so we'd have to increase our solar panel manufacturing by a factor of *54,800 times* - or a mere 5,480x if we were to accomplish this feat over a 10-year timespan.

As I said, difficult to take seriously.

At the end of the day, we will not be able to meet global energy demand with renewables alone absent a paradigm shift in renewable tech, especially since none of this math even includes storage. Whether it comes in base-load scale plants, or through commercial SMRs, nuclear is a vital component of the framework if we want to keep the lights on, or our atmosphere decarbonized. Unless of course fusion arrives, that is.