Hacker News Reader

Comment by sho_hn | Hacker News Reader

Can someone in the know summarize the improvements in this plant vs. the older crop? From the POV of "good ideas that made it to production", not vs. "what could be if only". What battles did these engineers pick and win?

View on HN

> A notable improvement of Gen III+ systems over second-generation designs is the incorporation in some designs of passive safety features that do not require active controls or operator intervention but instead rely on gravity or natural convection to mitigate the impact of abnormal events.

> Generation III+ reactors incorporate extra safety features to avoid the kind of disaster suffered at Fukushima in 2011. Generation III+ designs, passive safety, also known as passive cooling, requires no sustained operator action or electronic feedback to shut down the plant safely in the event of an emergency. Many of the Generation III+ nuclear reactors have a core catcher. If the fuel cladding and reactor vessel systems and associated piping become molten, corium will fall into a core catcher which holds the molten material and has the ability to cool it. This, in turn protects the final barrier, the containment building.

https://en.wikipedia.org/wiki/Generation_III_reactor#Develop...

The ahem warp core ejection system.

That's what Chernobyl had!

In Soviet Russia, reactor core ejects building!

(Couldn't resist. Accept downvotes. I'm sorry)

Honestly this is nostalgic. Reminds me of reading slashdot :)

It's certainly been a minute since early Slashdot. :)

The key innovation is ejecting the core downward into a contained vessel, instead of upward into the sky (to be pedantic, it was the lid of the chernobyl reactor that shot upward, the core subsequently started to burn).

>Many of the Generation III+ nuclear reactors have a core catcher. If the fuel cladding and reactor vessel systems and associated piping become molten, corium will fall into a core catcher which holds the molten material and has the ability to cool it.

Wow, that's actually pretty cool

I don't think it's true for the AP1000:

https://inis.iaea.org/collection/NCLCollectionStore/_Public/...

"Integrity of the Reactor Vessel is protected by surrounding it with water in the event of a threat of core melting, and therefore no core catcher is required"

In other words "we don't need a core catcher because we promise to keep refilling the boiled-off water after a blackout." There are other reactor designs that can safely shut down without any human action.

According to this, https://www.nrc.gov/docs/ML1117/ML11171A340.pdf, the two design alternatives are "dry cavity" (aka "core catcher") and "wet cavity". But the dry/wet distinction is a little misleading as according to this paper, https://www.kns.org/files/pre_paper/37/17S-854%EC%9D%B4%EC%A..., both alternatives require cooling water. The dry cavity design relies on indirect cooling--the water contacts the sacrificial layer ("catcher")--whereas in the wet cavity design the water directly contacts the core material, which has still effectively been "caught" in the cavity beneath the reactor vessel.

I'm not sure what all the pros and cons are for each approach, except that the wet cavity design has a higher risk of a steam explosion because of the direct water contact. But this seems to be addressed by containment structures designed for higher pressures.

EDIT: The succinct comparison of each approach is in the introduction to the second paper: "Some plants adopted the 'dry cavity' to enhance the spreading of the core melt on the cavity floor as well as to remove the steam explosion risk, while other plants use pre-flooding strategy to make the 'wet cavity' in order to enhance the coolability after RPV failure or to reduce the RPV failure probability."

These have more passive controls that don't require active management, and can go a few days with almost no pumps or anything running. They also have some improvements that avoid Fukushima-like events, with a core catcher that can catch and cool a molten core.

The biggest thing the AP1000 does is production scale with passive safety. With zero power and no operator intervention it can shut down a reactor and keep it cool for long enough not to melt down.

Is there a way an operator can command the operator into a failure mode?

By failure I assume you mean catestrophic disaster. The passive safety doesn't mean no failures, it just limits the damage done by them. The reactor is probably wrecked if the safety measures have to kick in.

But yes, if an operator (or more likely a whole shift of operators) was malicious, they probably could defeat the safety systems. The main innovation is basically just a huge tank of water that can drain into the reactor by gravity, so if you emptied this tank then the safety system is defeated. They could also just take a fuel rod, break it apart and dump it into a schools water tank. You can't really create a system that defends against murderous intent.

OK good point, but I'm guessing no one would be crazy enough to pick up a fuel rod and try and break it. I think you'd actually get burnt just from handling it.

In any case I was actually thinking about a foolish operator, not a malicious one.

Honestly only time will tell just how effective the passive safety here is.

Even with the previous generation of reactors it takes multiple failures all at once for an accident to happen, so who's to say one day we're not going to see some new unforseen issue happening just at the wrong time.

In fact this kind of touches on why some folks are looking at more radical changes. There are thorium reactor designs which require active, maintained energy to become critical at all. A melted block of thorium is sub-critical! None of these systems are any where near productionizable though.

Overall, remember this isn't the first time someone thought they finally "cracked" safety. This is really about us getting more experience and refining our practices over time, not some "now nuclear is totally safe" threshold.

https://how.complexsystems.fail/ was on hacker news a while back, you might like to read it to get more of a feel for how the progress of nuclear safety really happens.

From the manufacturer, Westinghouse, so please take with a critical look:

Key quote from the second link: "The key feature of the AP1000 plant is the replacement of complex redundant safety systems that are powered with AC power with passive safety methods such as gravity and heat transfer by conduction, convection and radiation.....

.....The AP1000 plant does not require AC electric power to achieve safe shutdown nor to establish and maintain, for an extended period of time, safe shutdown mode while removing decay heat from the nuclear fuel. By removing the reliance on AC power, you solve the paradox in which you need AC power to remove decay heat. With the AP1000 plant design, you don’t need AC power. You just need the laws of physics and stored energy from DC batteries, compressed gases and gravity to remove decay heat, and that is what achieves the simplicity and robustness."

Safety features overview page: https://www.westinghousenuclear.com/energy-systems/ap1000-pw...

Interviews with engineers for a salesy focused magazine (Not mobile friendly): https://digitaleditions.nuclearplantjournal.com/JA18/22/

Product overview page: https://www.westinghousenuclear.com/energy-systems/ap1000-pw...

It's an AP1000 PWR. Wikipedia has a summary of the design features:

https://en.m.wikipedia.org/wiki/AP1000#Design_specifications

Also adding onto this, what non-safety improvements were made? As I understand the PWR is a generally crap & outdated design that tends to suck up the benefits of nuclear in it's lackluster reliability and efficiency

This is not correct. PWRs are slightly less efficient than BWRs, but their safety systems are much more straightforward due to having an entirely non-nuclear secondary. PWR turbine halls do not need containment, whereas BWRs' do.

I suspect that this person wasn't talking about PWRs vs. BWRs, but rather, PWRs vs more-exotic gen-IV-ish designs (molten salt cooling, gaseous helium cooling, etc. etc.).

What is the biggest Gen-IV design ever built, in MW terms? Is it the 200MW Pebble Bed reactors in China that just started powering lights at the very end of last year? Given how difficult it has been to build the much more well-understood PWR reactors, I can understand the skepticism for trying to build a Gen-IV design.

Yeah, there are steep hurdles, but the hope eventually with at least some of the smaller gen IV designs is that a design can be approved once along with any requirements for siting, etc., and that then at each site where they're to be installed, installers will only need to show that the site meets the already-approved siting requirements rather than starting a whole new process from scratch, which should seems like it at least has the potential to make the regulatory burden more sane. The NRC is already working on a revised approval process that aims at this goal, and is set to be ready in 2024. Who knows if it'll pan out, but there's at least the possibility of change.

Reply on news.ycombinator.com