Hacker News new | past | comments | ask | show | jobs | submit login
Hyman Rickover on Nuclear Reactor Designs (1953) [pdf] (ecolo.org)
85 points by nkurz on March 29, 2020 | hide | past | favorite | 48 comments



Another ex-Navy nuke here. More than 45 years later, I can still practically recite both sides of my interview with the KOG,* which was necessary to get into The Program.

Here's an excerpt. Important to the story is that I was escorted into Rickover's office by a senior officer who was in training to be a submarine skipper. Also relevant is that I was younger than the usual candidate.

KOG: Why do you want in my program?

Me: From what I've seen of the submarine fleet, there's a lot of professionalism there; I'd like to be part of that. [Later, in nuclear power school, I switched to surface, and spent my sea time aboard the carrier USS Enterprise.]

KOG: So you don't think there's any professionalism in the surface fleet?

Me: I didn't say that, Admiral.

KOG [to the escort officer]: Read back what the kid said.

Escort officer: Mr. Toedt thinks there's a lot of professionalism in the submarine fleet and not in the surface fleet.

KOG: [Angrily points out my personal failings, ending with:] You're shooting your g--damn mouth off about something you don't know a g--damn thing about. What do you think of that?

Me: That's not what I said, Admiral. [At this point I started getting pissed off.]

KOG: [More vigorous imparting of wisdom, ending with:] Have you learned anything?

Me [angrier than I've ever been, before or since]: Yes, SIR.

KOG: What's that?

Me: Not to shoot my [pause] MOUTH off about something I don't know a [pause] thing ABOUT — SIR.

The interview went downhill from there, including inquiries about my sex life. I figured I'd bilged the interview and would be heading for destroyer school, which had been my preference anyway (the officers at my NROTC unit had strongly "encouraged" me to apply for The Program instead). I was shocked afterwards when the nice older lady congratulated me on having been accepted to The Program and asked which of the two nuclear power schools did I want to attend.

After the interview I realized that Rickover and the escort officer had intentionally been f--king with me as one of his little stress tests, the stories about which are legion [0].

* KOG = Kindly Old Gentleman. The nuke joke was that he checked the box for at most two out of those four things.

[0] https://bubbleheads.blogspot.com/2009/02/rickover-stories-ne...


I wonder what his reaction would be if you went with something similar to the Grand Budapest Hotel job interview scene.


You are likely familiar with the quip that "While in theory there is no difference between theory and practice, in practice there is." Here's a delightful letter by "The Father of the Nuclear Navy" Admiral Hyman Rickover (https://en.wikipedia.org/wiki/Hyman_G._Rickover) on how that saying applies to nuclear reactor design. Excerpts:

An academic reactor or reactor plant almost always has the following basic characteristics: (1) It is simple. (2) It is small. (3) It is cheap. (4) It is light. (5) It can be built very quickly. (6) It is very flexible in purpose ("omnibus reactor"). (7) Very little develop- ment is required. It will use mostly “off-the-shelf” components. (8) The reactor is in the study phase. It is not being built now.

On the other hand, a practical reactor plant can be distinguished by the following characteristics: (1) It is being built now. (2) It is behind schedule. (3) It is requiring an immense amount of development on apparently trivial items. Corrosion, in particular, is a problem. (4) It is very expensive. (5) It takes a long time to build because of the engineering development problems. (6) It is large. (7) It is heavy. (8) It is complicated.

It's a short two-page letter, so I won't excerpt more. It merits reading in full.


That really does sum up all the gaps between the promises of "too cheap to meter" and the practical expensive reality. Doubly so for things that haven't made it out of the lab at all, such as thorium reactors. There's always people willing to give the "academic reactor" sales pitch for nuclear on here, especially as a reason not to proceed with some scheme using a renewable, proven but intermittent technology like wind or solar.


He has a quote from somewhere that I can't locate at the moment which is (paraphrased) "any new reactor design is always safer and more efficient than existing designs, until they get built, which is when they become expensive and less safe".

I always have to refrain from trotting that out whenever the pro-nuclear fans trot out how wonderful Gen-IV, or MSR, or whatever lights up their fantasies will be.


My preferred formulation is “In theory, it works in practice; in practice, it works in theory.”


It's Rickover's concluding paragraph I find most compelling:

Yet it is incumbent on those in high places to make wise decisions and it is reasonable and important that the public be correctly informed. It is consequently incumbent on all of us to state the facts as forthrightly as possible.

That resounds today, especially with a novel current rapidly-spreading challenge, and responses around the planet which fail all three of Rickover's admonishments.

The essay itself has long resonated with me:

https://old.reddit.com/r/dredmorbius/comments/28q61o/hyman_g...


Off topic:

What are your thoughts about hoaxes, cults, groupthink?

I'm hoping to find explainers in the same way Sara Robinson and David Neiwert have tried to study and understand authoritarianism. https://dneiwert.blogspot.com/2006/08/cracks-in-wall-part-i-...

FWIW, I am scanning your prior posts (HN, reddit, mastodon).

I try to pay attention to "explainers" like you, Slate Star Codex, Gwern, Ezra Klein, David Roberts (also Vox), Bill Gates, others. People who voraciously consume info, synthesize it, and try to explain it to the rest of us.


That''s a very complex topic.

Of possible relevance, and reflecting one current of my thinking:

https://old.reddit.com/r/dredmorbius/comments/62uroa/clothin...


Prior navy Nuke- it’s hard to explain how much Rickover continues to permeate the Navy Nuclear program. When you start your naval nuclear career, you still literally have to go through Rickover. That is, the building where all naval nuclear training begins is “The Rickover” in Goose Greek, SC.


Rickover has a case to answer for the Thresher:

http://web.archive.org/web/20170514181829/http://www.au.af.m...

>Beneath the surface, however, was always the cold, unrelenting, ruthless workaholic, undermining the bureaucracy while creating his own. His was a management textbook for the inside operator, organized to the smallest detail, intolerant of error, devoting everything including his personal life to a cause, an obsession. Rickover established a constituency in Congress by superior salesmanship of his own product and skillful sowing of dissension and division in competing programs. Carefully slanting facts and covering up what he did not want disclosed, he skillfully manipulated his two bosses to become what the authors call "The Unaccountable Man." He destroyed any competing nuclear program within his own organization and any person likely to emerge as a competitor or successor. In time, he became increasingly conservative if not reactionary, putting space between himself and any responsibility for failure or accident. When the USS Thresher was lost in April 1963, he immediately phoned the Bureau of Ships to dissociate himself from any likelihood of failure of the nuclear plant in the incident. The bureau chief thought this action "thoroughly dishonest."


On the subject of this thread:

>The cult of personality produced other adverse side effects. Having achieved brilliant success with the pressurized water coolant system in the Nautilus installation, innovation in other types of plants was stifled. The USS Seawolf plant, developed in tandem with Nautilus, utilized liquid sodium as coolant, promising much smaller and more compact reactors. Because of limitations in metallurgy, the system was unsuccessful. The program was scrapped, and its obvious superiorities were never again reexamined, even after twenty years of further progress in nuclear technology. Rickover put the kiss of death on programs generated within the Office of Naval Research and elsewhere for smaller, lightweight reactors that could reduce the enormous size and cost of nuclear-powered ships. None saw the light of day; all were thwarted as interference in his work. When nuclear power was adapted to surface use for large combatants such as aircraft carriers and missile cruisers, new systems apparently were not examined. For instance, a smaller and more efficient combination of nuclear power for normal cruising plus an overdrive of conventional gas turbine plants for high speed use had been proposed but was not investigated further.


We built one. It didn’t work then. He said it wouldn’t, and indeed it didn’t. He’s been dead for decades, and we haven’t built a working one now—but somehow it’s his fault!


> Rickover has a case to answer for the Thresher

Why? The loss of the Thresher had nothing to do with any problem with its nuclear plant. It was a "subsafe" problem--inadequate QA on a pressure weld. The Navy responded with much stricter QA standards on pressure welds.


He made that phone call before he knew that, immediately upon hearing of the incident.


It doesn't appear to me that that affected the actual investigation or its conclusions, which, as I said, were that it was not a reactor issue but a subsafe issue. It's certainly relevant for the broader issue of how Rickover operated, but I don't see that it gives him anything to "answer" for the loss of the Thresher in particular.


Can you blame him?

Because rest assured, a lot of people would have tried to.


Wow! I wish the thorium enthusiasts would all read this.


Yes. Most of the problems with reactors have come from parts that are glossed over at the PowerPoint level. Many exotic reactor designs have been built, and most have had problems. Everything with complexity inside the radioactive area has been troublesome. The AVR pebble-bed reactor in Germany had a pebble jam, and still can't be decommissioned.

Often the problems involve plumbing. Sodium reactors have leaks and sodium fires. Helium gas-cooled reactors have helium leak problems. Ft. St. Vrain was a really good idea which had both helium leak and corrosion problems. Three Mile Island was a valve problem. Fukishima was a failure of the cooling system after a tsunami.

This is why some of the new designs are viewed skeptically. Molten salt reactors involve a chemical processing plant that works on radioactive materials. Radioactive chemical plants have always been troublesome. They tend to become toxic waste sites. (See Pantex, PUREX, etc.) Modular reactors with multiple reactors sharing the same cooling pool are at risk if anything leaks and contaminates the pool.

Commercial reactors have to be built for a 40-50 year lifetime to pay off. Anything inside the radioactive zone is essentially unmaintainable for that period. Complexity inside thus just doesn't work. Which is why we're stuck with water and pressurized water. Simple mechanically and chemically.


> Fukishima was a failure of the cooling system after a tsunami.

The fixes to that design would be straightforward. For example, putting the backup generators on a platform so they wouldn't get flooded.


The placement of generators at the base of the containment structures was already a mitigation to another risk: that ground movements common to the area might damage or destroy the generators.

A complex environment is one in which multiple risks constrain options and operations. Generators can be placed neither high nor low without assuming risks or requiring additional mitigations.

If a solution space is a search through a complex multidimensional topology, constraints place bounds on that space.


Come on, there's a lot of know-how to building a platform that can resist an earthquake.

An alternative is build a waterproof wall around it. Or enclose it in a waterproof enclosure. (We do know how to build submarines.)

With some thought, you can think of many more arrangements much more resilient than what they had, and would not have been particularly expensive.

Such as bulldoze a mound and put the generators on top of that.

Other problems with Fukushima had similar simple solutions. Instead of venting hydrogen gas into the enclosed building, where it built up and exploded, vent it to the outside. Cost: a vent pipe.


Not an expert in any of the related fields, but Japan is also a heavy seismic activity zone. Piling up loose dirt is just asking for liquefaction.

My armchair spectator opinion is that I want each reactor 'unit' built in an engineered safe solution, far up, supported by a seismic isolation system. Up in the hills, where people would be sent to evacuate after the earthquake due to the tsunami risk? Yeah, put that thing there.

Also the plant engineers need to be personally authorized and held responsible to do whatever it takes to keep things "safe" in the event of a disaster. They're on site, they're seeing the details, let them have the authority to ensure the greatest possible safety.


> Piling up loose dirt is just asking for liquefaction.

Having investigated this myself when I bought property for a house, this is not expensive to deal with. Heck, my neighbor did. You bore a hole with a big auger, fill it with concrete, and set the structure on it. You can even build skyscrapers on fill with such (see San Francisco).

The only reason these problems were not addressed when Fukushima was built is because people didn't think of it in their failure analysis, not because it is expensive, impractical, or impossible.

There is nothing fundamentally wrong with the Fukushima design, just details.


Up in the hills

That limits the cooling water supply. Almost all large reactors are near bodies of water. Palo Verde in Arizona is one of the very few exceptions. They operate a sewerage disposal plant to get water.


Small molten salt reactors were actually built. They weren't just paper reactors, even in Rickover's day. Now there are companies working on MSRs for commercial use. There's more computer simulation than running reactors so far, partly because governments are much more cautious now. But I'll also point out that Rickover in 1957 had no idea what computers in 2020 would be capable of.

I think it's reasonable to suspect that we didn't find the best possible reactor design six decades ago.


The part that the computers help with - nuclear and thermodynamic physics ODEs - isn't the hard part. A couple of sharp grad students can do that with OpenFOAM and a modest workstation.

The hard parts are the operations, maintenance, building for robustness, and enormous amounts of empirical field experience with the materials.

Rickover and his contemporaries judged - rightly - that those issues would make MSR and metal-cooled reactors too finicky to be of practical military value.


Maybe we shouldn't base our civilian energy structure on what has practical military value, even if it made sense as an expedient in 1957.

Operations is hardly an advantage of LWRs over MSRs, especially if you stick with simple designs instead of LFTR (as almost all real projects are doing). Those also make maintenance pretty simple. Building for robustness is a lot easier when you don't need build lots of active safety systems or an oversize containment dome due to high-pressure coolant. And some designs use existing nuclear-rated materials.


...but there actually was a liquid sodium reactor sub built in the 50s, from what I've heard.


Only one. She was a contemporary of the Nautilus. It was their experience with that vessel that killed liquid metal coolant for the navy.

It might help to understand something special about the nuclear navy in those days - prototypes in the lab don't count. Prototypes on land don't count either. The only prototypes that counted were ones that actually went to sea. All the way through the Skipjack generation, they were regularly doing crazy-eyed one-off experiments.

Nautilus and Seawolf were both prototypes.


Agree.

One Soviet 'November' class (K27) and all 7 'Alfa' class used liquid metal. It wasn't developed further or used after.


But Russia does have two sodium cooled reactors on their power grid right now.


And comining on sixty years later, still not decomissioned, because dealing with both chemically and radioactively volatile materials is a nightmare.


TIL there was a Seawolf sub long before the Seawolf-class. That explains a lot.


PBS had a good documentary on this.

https://youtu.be/G0p8bWYY7qM


My father-in-law was just getting used to his new posting when Admiral Rickover came for an inspection. Talk about pressure.


That culture of pressure never changed, and they are better for it.

The best analogy I can give you is this: Imagine your CTO just decides to walk up to your datacenter to throw open a breaker or few to verify that your failover and redundancy tech works correctly.

The CTO doesn't do this often - just once every few months. Each time the failure mode is different. Remove a live hard drive, disconnect some network switches, etc. The only reason she does it to is to emphasize the importance of these drills upon her VPs and directors.

They run drills of this nature on their teams every few days.


I actually had a boss who was a CTO who would do something like this.

He was a good boss.

At another job I did something daft and tried to recover a backup. The backups hadn't actually been recoverable for at least 2 years.


This happens to us regularly without someone having to go play chaos monkey.


Simulating a power failure to test and prove the resiliency of the system is precisely what caused the Chernobyl disaster.


Ah, no. The cause of Chernobyl is lost to the dead but probably caused by a young engineer who didn't quite do the right thing.

You absolutely must drill for rare, high stakes events. Auto-rotations in helicopters. Mass casualties in hospitals. Reactor scrams. Man overboard on a ship. All these things are drilled because you don't have time to read the book when it actually happens, and it rarely happens.


https://en.wikipedia.org/wiki/Chernobyl_disaster

The accident started during a safety test on an RBMK-type nuclear reactor, which was commonly used throughout the Soviet Union. The test was a simulation of an electrical power outage to aid the development of a safety procedure for maintaining reactor cooling water circulation until the back-up electrical generators could provide power. This gap was about one minute and had been identified as a potential safety problem that could cause the nuclear reactor core to overheat. It was hoped to prove that the residual rotational energy in a turbine generator could provide enough power to cover the gap. Three such tests had been conducted since 1982, but they had failed to provide a solution. On this fourth attempt, an unexpected 10-hour delay meant that an unprepared operating shift was on duty. During the planned decrease of reactor power in preparation for the electrical test, the power unexpectedly dropped to a near-zero level. The operators were able to only partially restore the specified test power, which put the reactor in a potentially unstable condition. This risk was not made evident in the operating instructions, so the operators proceeded with the electrical test. Upon test completion, the operators triggered a reactor shutdown, but a combination of unstable conditions and reactor design flaws caused an uncontrolled nuclear chain reaction instead.


WW2 Germany had Donitz, and Cold War US had Rickover - both were giants.

https://en.wikipedia.org/wiki/Karl_D%C3%B6nitz

Nuclear power is a game-changer, allowing aircraft carriers to sail completely around the globe at speeds fast enough to barefoot water-ski behind.


The words "nuclear" don't appear in this wikipedia article. It does state, however, that Dönitz was closely wedded to Nazi ideology.


You're missing the point.

Both were the successful leaders that their era and countries needed.

Donitz almost defeated Britain using his submarine fleet alone (he just didn't get the number requested.)

Rickover's accomplishment wasn't demonstrating that nuclear power was useful, it was getting the support and funding to actually roll out the new technology. To this day, no other country has a nuclear aircraft carrier besides the US and France.


75% of U-boat crew died in combat, primarily because Donitz never changed their cryptography and gave them tons of radio traffic to decipher.

"Radio traffic compromised his ciphers by giving the Allies more messages to work with. Furthermore, replies from the boats enabled the Allies to use direction finding (HF/DF, called "Huff-Duff") to locate a U-boat using its radio, track it and attack it.[64][65] The over-centralised command structure of BdU and its insistence on micro-managing every aspect of U-boat operations with endless signals provided the Allied navies with enormous intelligence"


They did change their cryptography. They changed the settings on the enigma machine daily, and they also went through various versions through the war.


Changing the settings wasn't enough after the enigma machine was cracked. And more to the point, Donitz management style made their radio traffic easier to decipher, and he was extremely slow to react when it was clear crews were dying because their radio traffic wasn't secure.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: