News for hackers. Stuff that matters.
Re:Public's Dread of Nuclear Power Limits Its Deployment by AnthonyMouse on Thursday May 16, 2019 @11:55PM (#19936116)
It's not about having no safety rules, it's about having more reasonable safety rules that weren't enacted with the goal of increasing costs.

Safety is always about trade offs. One solution makes things 30% safer and costs $2500, another makes things 0.000000002% safer and costs $2,500,000, another makes things 10% safer and costs $50,000 but is strictly an alternative to a newer technology that makes things just as safe but costs $500.

The problem with existing nuclear regulations is that they're stale and brittle. You're not allowed to do something which is a safety improvement because every change has to go through a slow, expensive bureaucratic process. By definition, preventing safety improvements does not improve safety.

But it also increases costs, because technology gets better over time in multiple dimensions, including efficiency. You can't use the technology which is both safer and less expensive because existing rules require you to do things in the strictly worse ways that existing reactors do them.

The process of improving the rules is itself the source of significant expense. Opponents will fight sound proposals on spurious grounds just to extend the duration of the process and increase costs. Or propose wastefully inefficient "safety improvements" that are politically difficult to argue against, because pointing out that you can get ten times the safety improvement at 10% of the cost with other technologies will only cause them to turn around and argue for all of those things too, so then your costs increase by 110% of the unreasonable amount instead of 100%.

The existing process is inherently broken. It isn't optimizing for safety or cost, both of which would be improved by a more efficient regulatory environment. Like the regulatory processes -- by no means perfect but a thousand times better than this -- that are used for other things with equally dangerous potential like dams and large chemical plants.

Re:Public's Dread of Nuclear Power Limits Its Deployment by pfdietz on Friday May 17, 2019 @07:44AM (#19938575)
And yet, you have not explained how reducing regulations is going to be done without also reducing safety.

Regulations get added because new accident scenarios come to light, often due to near misses in existing plants. So one would expect a regulatory ratchet effect over time, even with perfect rational regulation.

It's interesting to compare fission to another area of technology that has achieved high reliability in the face of accidents: commercial air travel. There, safety has been achieved on a pile of corpses. Aircraft are reliable because we've crashed so many of them over the decades and learned what not to do. And we can do that because society accepts the occasional smoking pile of aircraft and passenger fragments. And even there, when regulation eases off, safety suffers. Look at the recent experience at Boeing.

I don't think society would tolerate the string of meltdowns that would be needed for the same kind of natural reliability growth in nuclear.

Re:Public's Dread of Nuclear Power Limits Its Deployment by AnthonyMouse on Friday May 17, 2019 @06:54PM (#19944388)
> And yet, you have not explained how reducing regulations is going to be done without also reducing safety.

I have.

The existing regulations prohibit safer alternatives. The safer alternatives are also less complicated and correspondingly less expensive. Using those alternatives would not only not reduce safety, it would improve safety, but they can only be used if those regulations are removed.

Here is an example. Older reactors require active cooling. If they lose electrical power for circulating pumps, they melt down. This is precisely what happened in Fukushima.

Alternative reactors use passive cooling. If you have no electricity, it doesn't matter, having electricity is not required by design. Which means you don't need a bunch of expensive multiply-redundant systems and contingencies to keep the power on.

But the new designs aren't allowed by old regulations, so things are more expensive and less safe than they could be.

> Regulations get added because new accident scenarios come to light, often due to near misses in existing plants. So one would expect a regulatory ratchet effect over time, even with perfect rational regulation.

That is what often happens, but it isn't what ought to happen.

There are two ways to address a near miss.

One is to make the minimal change to the existing design to address the issue, which is the only real option for already-existing installations. So e.g. reactors that require active cooling should have better generator redundancy. The other is to fundamentally change the design so that that category of problem inherently cannot happen.

The second one is what you want for new builds, but it's also what a low quality regulatory response expressly prohibits, or at best makes unnecessarily more costly. Because you end up with a rule that says "must have double redundant backup power generation" even for a design that doesn't require any power generation.

All you need instead of that is the ability for regulators to make timely and reasonable decisions. If there is a rule requiring backup power for pumps then it shouldn't apply to designs that don't need pumps, and if that isn't clear then the process for determining that and correcting the regulatory error should take two weeks rather than ten years.

> It's interesting to compare fission to another area of technology that has achieved high reliability in the face of accidents: commercial air travel. There, safety has been achieved on a pile of corpses.

That isn't really necessary. When you need a system to be reliable you build it with high tolerances and significant redundancy of safety-critical components. Then failures come in the form of "the expected maximum value for this variable was 100, the material spec supports values up to 400, but we measured a value of 150 in practice." Then the response is to replace the material with one that supports values up to 600. But there was never any danger, only a smaller safety margin than anticipated -- which is what the safety margin is for.

This is dramatically more difficult for aircraft than almost anything else because an aircraft requires active systems to prevent it from crashing violently into the ground. The normal failsafe of "emergency shutdown in event of serious issues" doesn't exist because you can't just shut down a plane while it's flying through the air at five hundred miles an hour and thirty thousand feet. So the requirement there isn't just "don't blow up" it's that you can't actually stop operating no matter what happens or everybody dies. Which is a much harder target to hit. And yet they still kill fewer people per mile than cars by two orders of magnitude -- while traveling ten times faster and with a hundred times more passengers per vehicle.

We know how to engineer things to be safe without making them uneconomical.