This is about human psychology, rather than reliability in the engineering sense. I truly believe it would lead to better safety, but not perfect safety.

The idea is that yes, all or most of the engineers involved in a nuclear accident of any significance would kill themselves. Not only would their knowledge be lost, their fellow nuclear engineers would carve their names on a black stone, bury it in the earth beneath their corrupted stain, and never mention them again. Their families would be rewarded only if they showed honor by staying to fight the nuclear corruption, and dying honorably when the disaster is contained.

This won't work perfectly, but it would let society as a whole live with the contradiction inherent in nuclear power: it sustains us, while running the risk of turning our children's children into walking bags of cancerous mutation.

This is also pure thought experiment: what would it take to have a culture of safety as serious as the threat of nuclear disaster? I'm not willing to pay this price, and yet, I'm thoroughly familiar with the mass balance equations governing extractable energy and haven't a clue how to post-carbon without at least some nuclear.

But here's the thing: we have already had two "nuclear accidents of any significance": Three Mile Island and Chernobyl. Neither of those came even remotely close to "turning our children's children into walking bags of cancerous mutation". Nor will Fukushima. Yes, it's very bad for people near the plant, and for fishermen in the surrounding waters. But on a global scale, it's just not that big.

what would it take to have a culture of safety as serious as the threat of nuclear disaster?

Wrong question; the answer to it ask you ask it is that we already have a culture of safety at least that serious, with regard to nuclear power, because we treat nuclear power far more strictly than we treat other sources of power (like coal) that pose greater overall risk to humans. If we're going to effectively ban nuclear reactors because we're worried about radiation, we should also ban coal because we're worried about deaths from mining. And we should ban oil because of the risk of something like the Deepwater Horizon spill.

There isn't a simple solution to any of this; but a good start would be to be open and realistic about all the risks of all power sources.

Oh, I should have added: this claim isn't really true either. The WHO estimates that a million people die each year due to air pollution from coal plants. This is many orders of magnitude higher than the annual average deaths related to nuclear energy.

The difference is, there isn't a worldwide push to ban coal because of this; instead, there's a push to scrub the plant emissions to remove the pollution. So again, nuclear is treated much more strictly than other power sources, relative to its actual risk.

I don't believe so. Toxicity and Emissions of a coal-fire power plant does severely affect those in the neighborhood. Studies show that radiation dose is comparable. As and Se are found to have higher concentrations in soil samples.

http://rpd.oxfordjournals.org/content/18/3/169

So killing or having adverse health effects on far more people is OK, as long as the effects are diffuse? I'm not following the logic here. The number of people affected by Fukushima, even on the worst-case estimates, is less than the number of people affected in a year by the various aspects of coal plants that posters here have mentioned.

And I'm saying that this local perspective (which I understand might not be yours, you might just be trying to articulate the perspective of the NIMBY folks) is flawed, because it doesn't treat equal impacts on self-interest equally. It protests against nuclear plants, while not protesting at all against coal plants that have a greater impact on the same person's self-interest, simply because the impact of the coal plant is "diffuse": it pollutes their lungs over time, all the time, instead of bringing some probability of being in a Fukushima zone, but the total impact on expected years of life (or expected quality of life) for that person is still greater for the coal plant.

I understand that this perspective is common, but that doesn't make it right.

Furthermore, the risks of emphysema and cancer due to coal pollution are underrated. Coal causes much more deaths among people who don't live in coal plants than we like to pretend. (Conversely, nuclear is safer on the balance.)

No, it doesn't. It's a difference between risks taken willingly by the people who benefit from the outcome, and risks imposed on people who don't have a choice and don't stand to benefit. And it's obvious that it was meant that way. Don't be deliberately obtuse.

The largest part of that mercury comes from coal-burning power plants. Coal power has contaminated the entire planet to the point where any food from the ocean has to be limited for vulnerable people.

There is no way you can possibly say that coal doesn't impose risks on people who don't have a choice and don't stand to benefit.

Coal is far more dangerous than nuclear by nearly every measure. The only reason it's more generally accepted is because that danger is spread out over a much wider area. Global coal power is analogous to a constant, ongoing Chernobyl, except instead of covering a small chunk of Europe, the same effects are diluted over the entire planet.

We all pay a measurable and fairly significant price for coal power generation in terms of shortened lives and increased chances of various disease. Nuclear power has not contaminated the planet to even remotely the same degree.

I personally think nuclear power can be made relatively safe, and that fears about accidents are exaggerated. I welcome disagreement there, but it's not possible to make a charitable comparison with coal in any way. The disparity in opinions and dangers between nuclear and coal is very real.

Huh? People do have a choice to move if they don't want to live near a power plant. And the plant is making the electricity they use, so they do benefit from it.

(That said, plenty of nuclear plants have been killed because the people who would have lived near the plant protested.)

No, the poorer you are the less attractive choices you have. But you still have the ability to weigh your choices and pick the one that poses you the least risk, all things considered.

The problem is that you have no idea how Process Safety works, or even what it is, judging by your posts. As the parent to this post explained, these major hazard events are probabilistic, not deterministic. That means that they can never be completely eliminated, only guarded against and reduced to the point that the benefit is understood to outweigh the risk. There will always be risk.

You suggest that the engineers responsible for designing the plant should commit suicide if this once in a million year event takes place. What about the businessmen who accepted the risk? Or the government who stipulated what the tolerable risk was? Or the people who moved nearby, aware of the risk? Or the workers on the plant who, knowing what the correct maintenance procedures were to keep the plant safely operational, ignored them because it was 'too hard'? What if the root cause was a vector of risk which was poorly understood by everyone at the time? Or the engineering specification was correct, but the materials vendor skimped on QA? Or were lied to by one of their plants? Or a subcontractor of the plant? Etc etc etc.

What you propose, even as a thought experiment, is futile. Because even if you could reduce the risk to only systematic errors made in engineering design, if there was a perceived need or benefit to design the plant, you could always find a group of engineers desperate enough, naive enough, or vain enough to take on the risk. And therefore society as a whole would not be safer.

As it happens there's a large body of engineering knowledge which goes into making dangerous plants 'safe enough' to outweigh the risks to their workers, owners, and society at large. It's not foolproof, but arguably it does a good job of allowing societies to benefit from the dangerous processes we seem so desperate to use. And don't kid yourself - just because these massive events incur a media feeding frenzy, doesn't mean smaller hazardous occurrences don't affect the lives of thousands of people every year. Should we kill everyone who could in some way be held accountable for another's death or injury?

In short, don't let fear guide you into silly knee-jerk reactions. There are massive problems inherent with dangerous processes, and most of them relate to the very human failures of laziness and greed. We're working to improve the situation constantly, but it's a long road and a constant battle.

I'm not advocating ritual suicide, but I think it's worth noting that in the case of Fukushima, there is an easily identifiable single root cause to all the current problems: the fact that the backup diesel generators and the switchgear for them were sited behind a seawall that was overcome by the tsunami. If they had been on higher ground, none of this would have happened; there would have been a continuous supply of cooling water to the reactor cores and things would have shut down normally. The process that led to that siting of the backup generators should be fairly easily traceable, as compared to the design process for the entire reactor complex.

Sure, but who's responsible for that? I don't mean to say that there must be many technical causes of incidents, but rather that it is almost impossible to find a situation where there is a singular human cause.

The multitude of engineers that saw the plans before they went final? The managers who pushed for the generators to be placed there in the first place? The contractors who built it knowing full well that the design was garbage? How high up does it go, and how many degrees removed from the situation do we require before you're absolved of guilt?

We could do something silly to consolidate the guilt, like, "the last person to stamp the designs gets the blame", but that seems like it would encourage a culture of hot-potatoing instead, especially if the stakes are your own life ;)

To put a more extreme example of it: a bus driver falls asleep at the wheel, jumps a curb, and kills a person. Do we punish the driver for his negligence, or his shift supervisor for threatening him into taking an extra-long shift, or someone even further up who knowingly encouraged a culture of overwork?

But if you can focus in on a single technical decision that had such a large impact, you can at least focus attention on how that particular technical decision got made. Yes, many humans will have been involved even in that single technical decision; but it's still a lot more manageable to look into how it got made than to look at how the entire plant got designed.

We could do something silly to consolidate the guilt

I'm not talking about guilt; I said I wasn't advocating ritual suicide. I'm talking about how to prevent it from happening again. To do that, you need to be able to focus on something that can actually be changed. Take your bus driver example:

Do we punish the driver for his negligence, or his shift supervisor for threatening him into taking an extra-long shift, or someone even further up who knowingly encouraged a culture of overwork?

Depends on the facts. Was the driver negligent? Could he have reasonably predicted that he would fall asleep at the wheel? How could he have acted on such a prediction?

Did the supervisor actually threaten him into taking an extra-long shift? If so, how did he get away with it? Aren't there rules about preventing driver fatigue? (Certainly commercial airline crews have strict rules about how long they can fly before they have to rest.) If not, why not?

Is there a systemic culture of overwork, encouraged from higher up? Do bus drivers in this company routinely drive longer hours than the industry average? Than the average in other similar occupations?

The point is that if you can focus in on specific causal chains, you can direct efforts at changing them. In the case of Fukushima, if you can focus in on how the technical decision to site the backup generators that way got made, you can direct efforts at improving that process. Did the engineers just not realize that siting the backup generators that way was a bad idea? Was the question never even asked? Or did the engineers realize it, but management overruled them? Or did the original design have them sited differently, but time pressure during construction persuaded someone (not the engineers or managers who did the original design) to change the location in order to meet some deadline?

All this, btw, reinforces your point that ritual suicide doesn't help anything.

Such things sometimes can't be predicted. If you fall asleep because of exhaustion, you don't necessarily see it coming (it happened to me several years back, one minute I was fine and the next I blacked out). You cannot criminally persecute a person for falling asleep. Only if he did something against the law, like also being under the influence of alcohol.

> Did the supervisor actually threaten him into taking an extra-long shift?

You know very well that supervisors rarely have to threaten their subordinates in low-end jobs. I have a personal acquaintance who's a truck driver. And he was explaining to me how when the new management came in, they first did some lay-offs based on silly reasons and after that nobody dared arguing.

Also do some reading on Japanese culture. Subordinates rarely question their bosses (or change jobs for that matter). It happens in many cultures. If you're from the US, then what's natural to you may not apply in other countries or cultures.

I'm not saying that blame cannot be assigned (it can clearly be, especially in the case of broken engineering designs), but seeing people here suggesting "ritual suicide"? OMG.

But why would you even take that risk? Not to mention that tsunamis can easily reach 100 feet in wave height; I'm not sure you could build a seawall that could keep one out.

Besides, it was pretty clear to me reading the sequence of events at Fukushima that there was entirely too little thought given to backup decoupling - there was too much of a zipper effect of one failure leading to another.

I'm not a lay person on this - I worked for 3 years on airliner designs, and much of that is finding ways to deal with failure cases. The nuke industry could learn a lot by consulting with airframe engineers.

Yes, that would work, as long as you also protected the switchgear and the power cabling leading to the pumps that serve the reactor.

The nuke industry could learn a lot by consulting with airframe engineers.

I think the nuke industry in general does pay attention to these things; I was actually somewhat surprised to find that the designers of Fukushima had missed such an obvious failure mode.

1. hydrogen being vented into an enclosed area

2. no backup method for determining coolant levels

3. backup generators were not protected

4. no backup method for adding coolant (I would have had a gravity fed backup system)

5. critical systems were located too close to the reactor - I would have moved them further away so they could be repaired without the workers being irradiated

None of these are expensive to do.

3. Backup generators were located in different areas. The only mistake is that they were ALL at ground level. That's unfortunate that they were all wiped out at the same time but one of the largest Tsunami ever, some stuff that almost never happens. They'll learn from that.

> None of these are expensive to do.

Yeah, when you are building the plant, it's not expensive to do. When you have to modify an existing design, it's way more expensive and you'll have to make a good case for why you need it.

The key idea is nobody has managed to design parts that won't fail. But they have figured out how to design highly reliable systems out of parts that fail.

Current designs use lots of redundant safety systems. New designs depend more on simple physics...eg., fuel that expands enough when it heats to damp down the reaction, or large quantities of coolant with very high thermal capacity and conductivity. At Argonne's IFR project, they shut down coolant circulation and electric power, and the reactor quietly shut down with no damage.

The ultimate solution is probably the molten salt reactor. With liquid fuel you continually remove fission products, and with most of those gone you don't have decay heat issues anymore.

What I would like to see is mandatory escrow of a large percentage of profits from highly-profitable, possibly-disaster-prone industries. If you go disaster-free for 20 years you get your escrowed money with interest. If not, you lose (an appropriate amount of) it.