Nuclear regulation now should be more like pipe safety regulation during World War II

Our civilization would be better off if it had dramatically more, cheaper, cleaner energy. To achieve this, we should allow nuclear power to become dramatically cheaper and more common, and the main barrier to this is over-the-top safety regulation. The history of the ASME piping and pressure vessel codes, which govern both oil refineries (my expertise) and nuclear plants, offers an important but generally unknown precedent for dramatically rationalizing so-called safety standards. The lesson can and should be applied to current nuclear regulation.

Long story short, pre-World War II, heavy industry was coasting along using a piping safety design philosophy that was well-proven but (as it turned out) excessively conservative, and therefore needlessly expensive. The resulting costs were passed to consumers, and no individual had an overwhelming incentive to rock the boat. Then,

In August 1942, the War Production Board requested a review of measures to conserve vital materials in piping components. A special War Committee of B16 was appointed and, operating under War Standard Procedure, developed revised pressure-temperature ratings for all materials and pressure classes.

Which is to say, America decided to man up and cut some of the fat out of pipe safety ratings. Pre-war, a nominal 300-pound pipe flange—the rim on the end of a pipe that allows it to be bolted up to something—was indeed considered good to 300psi, at 750 deg F. The wartime ratings increased that to 425 psi. Performance increased 42% at the stroke of a pen because society felt the urgency of the situation, executive authority in the form of the War Production Board threw down the gauntlet to the engineering community, and the engineers rose to the occasion. “Their goal was to make Carbon Steel flanges more acceptable for the defense industry projects during WWII without relying on alloy steel flanges,” freeing alloy elements for other war use.[1]

This reduced the factor of safety in flange design. This is more or less the same thing we’re too terrified to do in nuclear engineering. But the original flanges were so over-designed that the reduction in safety factor made no difference – there was no rash of flange failures beyond the base rate. (There are many reasons that flanges tend to leak at some base rate, regardless of pressure rating, e.g. assembly-related issues.) And the few engineers who really understood flange design could have predicted this outcome. The new ratings were still asymptotically safe, close enough to the highest achievable standard of safety for any practical mortal purpose. So they were permanently adopted after the war, and similar ratings remain in use today.

To illustrate asymptotic safety, think of standing in front of a firing squad, inside a protective steel box. If a bullet penetrates the box, you die, but a little bit of steel box goes a long way. If you have a four-foot-thick steel box, and you reduce its thickness to one foot, you reduced the “safety factor” by 4x. But your risk remains asymptotically close to zero. The bullet simply will not make it through. In fact, you might be better off in the one-foot box, when the lid gets stuck. And anyone with industrial experience knows that the lid is probably going to get stuck—human nature bends events towards unexpected perversity, and lazy overdesign can create problems elegance would never dream of.

Also during WWII, the following was published in the 1943 ASME “Unfired Pressure Vessels” Code:

These rules in their present form do not include any consideration of casting quality factors, weld metal efficiencies, and relative effectiveness of design details. The proper application of these factors is being studied by the Committee and, in view of the war emergency, is left for the consideration of the designer for the time being.

This shows clear recognition that the nation faces a crisis, the de facto regulator[2] doesn’t have all the answers, and so engineers ought to be free do something reasonable and keep things moving! If only our medical regulatory system took this approach! Instead, we have the Pointy-Haired Boss – “[PHB:] I’m afraid to make decisions because I might make the wrong ones. So instead, I tell my staff I need more data. [Catbert:] When really you need more brains and courage.”

So what could we “[leave] for the consideration of the designer” “in view of the [reported climate] emergency?”

In nuclear engineering the problem isn’t flange design or casting quality factors – it is the death-by-a-thousand-cuts regulatory framework built up over decades of a safety ratchet effect, where every bright idea anyone ever had for making a nuclear plant safer ended up mandatory, regardless of cost or future technological advances, for all subsequent construction. Thus nuclear has a wrong-direction learning curve: instead of getting cheaper as we build more, like solar panels, batteries, integrated circuits, etc., it has become more expensive (at least in the U.S.). There is no single number or requirement that can be changed to solve the problem; the bulk of the system needs to be thrown out and replaced. I refer the reader to the excellent “Why Nuclear Power Has Been a Flop” book by Jack Devanney, or this review and summary of it, for details of what’s wrong and some specific policy suggestions.

And if we want much cheaper or much cleaner power, slashing nuclear regulation is the low-hanging fruit. Natural gas and coal are already about as cheap as a relatively unregulated market can make them. Both likely cause far more deaths than nuclear. Hydro is limited by geography, and wind and solar need backup by base-load or (expensive) battery power. All of these avenues are already receiving billions in investment, while we basically dropped the ball on nuclear two generations ago and haven’t had the will to build a commercial-scale new-technology reactor since.

If we were serious about either climate change or economic growth, surely we would resurrect nuclear progress by taking something like the World War II approach. The high political authorities concerned could confront the engineering community with the ultimatum: “We need more nuclear power, and we expect and require actual engineering progress—progress in elegance, simplification, and above all cost—even at the cost of an increase in perceived and perhaps even actual risk.” Brave and public-spirited engineers, with adequate backing, could take responsibility for a blank-slate safety regime and pushing the field beyond the stale, comfortable status quo. We would aggressively test (not just interminably model) new full-scale reactor designs—at remote sites initially, but nonetheless actually testing, so as to learn quickly and promote long-run safety. (If you have a hundred million dollars and/or political influence in Indonesia, Devanney’s Thorcon company is working along these lines and would likely appreciate your help.) We would also face and accept the irreducible risks of nuclear power, which are higher than the U. S. regulatory framework pretends, but lower than the public believes, and lower than the fossil fuel alternatives (see Devanney’s book again).

If I was emperor, my cartoon policy[3] would be to appoint Elon Musk king and Jack Devanney’s team dukes of a remote pacific atoll, with exemption from all current nuclear and environmental regulation, ten billion dollars, and orders to go and achieve a continuous year of full-power operation on a gigawatt-scale molten salt reactor before coming back. I’d also give them a ten-pack of “get out of jail free” cards for melting down floating reactors and sinking them offshore—water is an excellent radiation sink, and indeed plenty of nuclear waste has already been safely disposed of in the ocean. Maybe Musk could just go back to Kwajalein, which is where, for similar reasons, SpaceX had to start orbital launches. (I would even be willing to give Jeff Bezos another atoll, and tell him to get to work too.)

Within a realistically constrained policy option space, I am not terribly optimistic for our odds of doing anything like this. Nuclear regulation is just one more way in which Our Regulatory State Is Failing Us, and if we can’t make reasonable risk-based decisions when we are in the midst of a global pandemic (e.g. we obviously should have had vaccine challenge trials during Spring 2020), I don’t like our odds of making wise tradeoffs, against possibly greater voter prejudice, for payoffs further in the future. We are so absurdly unserious that we can’t even build long distance transmission lines.

Nonetheless, I wish I could draw society towards a bold commitment to cheap nuclear power. I’ve lived in Brazil as a missionary for the Church of Jesus Christ of Latter-day Saints[4], at something like 1/8 the GDP per capita I’m used to, and I’m here to tell you that reducing energy use that much is miserable. I really, really value my air conditioning and my clothes dryer. Symmetrically, increasing energy access could yet make my spoiled American life better; energy costs are still too high for me to have an indoor ski hill nearby. And since energy is a major component of costs for food, steel, concrete, manufacturing, transportation, and thus practically every physical convenience of modernity, making it cheaper will improve the productivity of the entire economy. And I guess cheap nuclear power could also benefit countless people in the third world and dramatically reduce anthropogenic CO2 emissions at an unsubsidized economic benefit by outcompeting coal and natural gas generation.

Nuclear engineering is currently limited by courage, political will, and the status quo, not by scientific knowledge. Like during World War II, political and thought leaders should give courageous engineers the cover they need to cut the fat, giving up some trappings of safety, in order to allow progress that can later be recognized as an unalloyed good.


1. This quote and these pressure ratings are from the March 13, 2017 agenda of the ASME B16 Subcommittee C on steel flanges.

2. ASME is a private organization, and much of the time, much of the code is written by industry volunteers. But most states (though not all) have come to legally require code compliance, and even in states (like Texas) where code compliance is not specifically required, common sense generally leads people to comply anyway. There is also a national regulatory requirement to use RAGAGEP (Recognized And Generally Accepted Good Engineering Practices) that would generally result in ASME code compliance even absent state mandates. 

3. A cartoon policy is like a cartoon mathematical model – more fun to work with than the full-blown version, and comparably illustrative. Certainly, though, a successful effort ought to have a strong, engineering-obsessed, unitary authority in charge (which Musk provides in his companies), and readiness to take risks and test real hardware, potentially to failure (ditto).

4. Which church, by the way, is underrated, worthy of your attention, and true. I served in São Paulo and environs 2010-12.  

Pipe flange featured image courtesy of Wikimedia user Kkmurray.

Write a Comment


  1. Well written and thought out, Tom. I have also long thought that we are completely neglecting the potential for safe, reliable nuclear power. As an engineer from a past era, I can tell you that ideas like you propose, are needed now and would be far more effective and fair to all countries in addressing emissions than the airy-fairy carbon taxes and regulatory quagmires.