"Nothing like this will be built again": a tour of a working nuclear reactor | Hacker News Reader

"Nothing like this will be built again": a tour of a working nuclear reactor

antipope.org

160 pointsgjm1137 comments

When I worked in academia I did one project with Scottish Nuclear and one of their team had worked on the project to build Torness. He had a story about one of the failure modes they had to consider during the design.

AGRs can be refueled while they are operational by a huge machine that moves across the top of the reactor and allows the fuel rods to be extracted from the extremely high pressure environment of the reactor core. The refueling machine is basically a long thing container that fuel rods can be raised into or lowered from.

They had to consider the case of what would happen if the refueling machine wasn't properly pressurized before the connection to the reactor core was opened (which can't really happen for a variety of reasons). However, if it did happen they realized that the refueling machine would basically act like a very large gun shooting hot fuel rods through the roof of the reactor building powered by the high pressure CO2 from the reactor core.

Every time I pass Torness on the train I can't help looking to see if there are any fuel rods flying into the air. So far, so good....

"Energy to cheap to meter". It pains me. I'm so glad to see a good writer give some strength to how massive and impressive these reactors are. But it pains me as he describes system after thoroughly redundant system adding to the costs of these reactors.

I'm much more familiar with PWRs (Pressurized Water Reactors) which yes, are a bit smaller and less efficient, but the plans and schematics are like a first time run through of a problem. Nothing re-factored, nothing elegant, just plain and simple brute force design to make it work.

There's no reason for four diesel generators when two would do. There's no reason for multiple chemical volume and control systems. There's no reason for 6 foot thick, steel re-barred concrete containment in a negative void coefficient reactor!

My take-away from working in the industry is regulation severely impedes innovation. I fear that our greatest discovery since fire is languishing. Nuclear energy will be used to great extent in the future, but I can't stand the current delays. Energy is conceivably too cheap to meter in it's current form. You could conceivably pay, not by your kilowatt-hour, but a fixed monthly subscription to unlimited energy. Just think about what you could do with that!

Edit: -- Misread the article. PWR's are less efficient than AGRs

Sorry, they are less efficient not more. Getting 43% from a nuclear power plant is incredibly difficult.

Noted and edited. However, not the point of my post. AGR's, PWR's, and all current reactors for that matter are grossly over designed.

43% of energy is incredible but the redundant systems of the plant have nothing to do with the efficiency of the plant.

Pebble-bed reactors can approach 50% thermal efficiency.

Have we actually built one that gets this kind of efficiency? I've heard people saying that pebble beds could combine high core temperatures with gas turbines to get around 50% efficiency, but the only commercial-scale pebble beds I know of are the Chinese HTR-PM, which uses steam turbines to get around 40% efficiency. Still damn impressive, of course.

Here in Canada they shut down the worlds major reactor for producing medical isotopes - because the secondary backup diesel generator failed an inspection.

It didn't matter how many cancer patients died because of the lack of treatments - you can't be too careful.

It's lack grounding all the fire engines in a city because one of them had a siren a 1/4 tone flat.

Actually, Chalk River went down for awhile due to a crack in the reactor vessel. It's actually rather amazing that they were able to repair it. They actually had to use "robots" to repair the welds because it was too hot for humans. The imaging techniques (effectively sonar and eddy currents) that they used to determine where the fractures were was also impressive.

So, sometimes if something shuts down, it's not a conspiracy, but really is an issue of safety (cracked reactor vessel===bad)

>There's no reason for 6 foot thick, steel re-barred concrete containment in a negative void coefficient reactor!

To contain the steam explosion in the event of a coolant leak. That is, pressurized liquid water at 300 °C flashing into steam as it's depressurized.

Coolant leak and steam explosion are in conflict with one a other. What you're referring to is Chernobyl, which had a graphite moderator which allowed for a dramatic increase in power and pressure with no release. Light water plants use water as a moderator and a coolant leak would relieve pressure making a steam explosion unlikely. For example Three mile Island was never in danger of a steam explosion.

>Light water plants use water as a moderator and a coolant leak would relieve pressure making a steam explosion unlikely.

I mean, the pressure in the containment building. Isn't that the purpose of the containment walls, to withstand steam pressure from a coolant leak? Or am I misunderstanding things.

>Coolant leak and steam explosion are in conflict with one a other.

My mistake.

My argument is that there is no purpose for the containment walls. They are theoretically yet another barrier to leak of radiation on top of the fuel rods and the piping itself but I would argue this as unnecessary. Chernobyl should have had a containment vessel. This would have averted that disaster. Three Mile Island had a catastrophic systems failure with a partial meltdown yet only 1 inch of the 7 inch thick reactor vessel was affected. Some xenon gas and other radioactive elements were released to the atmosphere but of minimal quantities.

Isn't this a gas cooled reactor?

Yes, the parent brought up PWRs.

My take-away from working in the industry is regulation severely impedes innovation.

Regulation is a lousy form of idiot proofing in terms of efficiency. But it's somewhat reliable.

What do you with a technology that's profoundly useful and perfectly safe in the hands of a sane, reasonably intelligent person but grotesquely dangerous when in the hands of a greedy idiot?? Don't tell me there aren't lots of greedy idiots around...

The inability of biological science to freely perform experiments on humans really does hold back a lot of potential innovation also. That sounds snide but it's also literally true.

Science and society have an uneasy risk-reward relationship.

This is true. Given that there is no shortage of examples of harmful mishaps in the nuclear industry caused by corner-cutting and mismanagement, the mind truly boggles at what might happen if regulation was any looser.

There's no reason for four diesel generators when two would do. There's no reason for multiple chemical volume and control systems.

So you are saying this reactor type is still safe even when both diesels fail or when the chemical volume/control systems fail?

I'm honestly curious (my knowledge about nuclear reactors is fairly minimal).

"There's no reason for four diesel generators when two would do"

I remember one senior Scottish Nuclear bloke telling me about one, rather important, bit of equipment at Hunterston B plant (another AGR site on the west coast of Scotland) - the Reactor Shutdown Safety Sequencing Equipment (RSSSE) (not my recall of the details of this acronym are probably a bit off). This shuts down the plant in a safe way if something were to go badly wrong.

Apparently when the AGRs were being designed in the 60s/70s they, understandably, to use tried and tested technology for this module. However, what tried and tested meant to these engineers was that stuff should have had multiple decades of production use before they would consider it - so the technology in the RSSSE was essentially from the 1940s - worm gears, relays etc. At the time I was working on research projects that involved Scottish Nuclear (the early 1990s) they were rather proud of the fact that there was no safety critical software in any AGR plant (there were plenty safety related systems).

So I think it's safe to say that the designs of these plants were, when it came to safety, quite conservative - even if the overall design goals of the plants were actually quite ambitious. As Charlie says, the abiding impression I got of the design of the AGRs (and I am no nuclear engineers) is that they had a design decision that would impact safety they always played safe - massive redundancy in systems, a vast amount of instrumentation (my estimate of the number of signals fed into the control room was off by a comical factor) etc.

We have in Scotland an excellent example of what could be called "conspicuous over engineering" - the Forth Rail Bridge - which was designed in the aftermath of a rail disaster where a bridge over the Firth of Tay collapsed with a train on it, causing great loss of life. With the Forth Rail Bridge not only is it a fantastic bit of engineering the whole spirit of the thing asserts "this is safe".

To me, my experience with the people of Scottish Nuclear and the design of the AGRs themselves gave me a huge amount of confidence that not only were the designs probably as safe as they could be but they were operated in a way that was quite amazingly safety conscious. Given that the UK is rather small and very crowded - with the Central Belt of Scotland being particularly densely populated and neatly bracketed by AGR plants at Hunterston in the west and Torness in the East I'm actually rather pleased that these systems were also conspicuously over engineered.

They could have used some more of that at Windscale.

Actually, I did wonder whether there might have been a backlash against the crazy under-engineered Windscale air cooled reactors and the near disaster of the reactor fire there:

http://en.wikipedia.org/wiki/Windscale_fire

So Windscale might have been the Tay Bridge to the Forth Bridge of the AGRs.

[Edit: The description of the fire in that Wikipedia page is pretty alarming - fancy sticking a scaffolding pole into an on-fire nuclear reactor and bringing it out dripping in liquid irradiated uranium!]

When I was getting my degree is Physics we had a lab in the "student reactor", which was still in a large building of its own. As part of the lab we got to sit down to the control console of the reactor and control it, on our own. I remember I was turning one knob increasing the power output when suddenly there was a loud noise, everything shook, and the power output went to zero. Turned out it was the failsafe (I think cadmium rods dropped into the core) as I was too agressive with the controls --- it was a normal occurence for students. (All modern reactors are designed in a way that you can't overdrive them, they'll automatically shut down.) It was great fun and a story I'll be telling for the rest of my life.

> For starters, some embedded controllers in racks in the auxilliary deisel generator control rooms have EPROMs which have been known to be erased by camera flashes in the past

Not being in EE, I didn't know that this was possible; and, with the knowledge that a bright flash erases EPROMs, that they'd be left in a position where this could occur. More interestingly, the NRO has a notice about it: http://www.nrc.gov/reading-rm/doc-collections/gen-comm/info-...

The government verbiage for the hack here (tape over the window) is amusing in itself:

> They also confirmed that the light from the Canon flash and the Polaroid flashbulb could be effectively blocked by "black bagging" the flash, or by blocking the EPROM window with "tin foil" held in place by clear cellulose tape, or by blocking the EPROM window with "standard electrical tape."

I remember seeing a motherboard or video card BIOS stored on EPROMs as a kid, and wondering what it would happen if I took the light-blocking sticker off and shined a black light in the computer while it was running. It is interesting, though not too surprising, that the EPROMs are exposed in the racks. If nobody's ever going to be in the room, why bother putting walls on the racks?

Also, it's probably a good thing that flash memory was invented before the ultraviolet LED, otherwise we might be stuck with LED-based EPROMs with painfully slow erase cycles ;).

Trident 8800CS! Had hi/lo eeproms with a reflective hologram sticker over the eeprom window. Now we upload microcode/firmware on bootup. We have things so good.

I loved my Trident. It was the first video card I owned that could do 1024x768. It's been so many years that I don't remember all the other steps I went through between Hercules and CGA to a modern nVidia, but almost every one of them was special in some way.

Really? I hated my trident. It was dog slow. Now, I loved my tseng et4000. Fast, great mode-x and tweakmode support, easy to code for.

Since I was rather young at the time, my access to anything better was limited. Talking about mode X does make me nostalgic, though I used Allegro instead of writing my own low-level init code.

Not being in EE, I didn't know that this was possible

Not only is it possible, it has led to many interesting failure modes in the past since visible light doesn't usually erase an EPROM, but sometimes just flips a few bits. The n00b complaint, "it works fine until I turn on my desk lamp" used to be a pretty common one in EE labs and online help forums :-)

I don't buy it for a minute. You can leave an uncovered EPROM in direct sunlight for a week, and the data will probably survive.

I'd suspect a localized "EMP" transient from the camera flash before I'd suspect an erased EPROM. It takes a lot of energy, in the E=hv sense, to erase an EPROM.

To actually erase one, yes. But to flip a bit that's on the edge, not that much.

Believe it: windowed EPROMS have been used as cameras. Their light sensitivity is legend.

[edit] actually I was wrong about that. The memory devices used as cameras were delidded RAMs, not EPROM. EPROMs are light sensitive, but too slow for that.

You used to erase eeprom with UV lights. I remember doing this back int he late 80's early 90's in my dads basement electronics lab. Fun stuff. Now we just flash everything with software. We have it much better these days.

This has been posted here before, about two years ago -- http://news.ycombinator.com/item?id=569564 -- but (1) it's very good and (2) the old URL no longer works.

It's by Charlie Stross, a science fiction author and former hacker. It describes a lengthy tour of the nuclear reactor complex at Torness in Scotland.

Not to mention that it's an AGR (Advanced Gas cooled Reactor) something that doesn't exist in the U.S. or most of the world. Very cool.

Very clever design, much more efficent, much safer - negative temp coeff and no chance of a steam explosion.

But they never reached the critical mass (sorry) to make them standard and so cheap to build and operate.

He also posts on HN as 'cstross'.

Here's a large (35 inches) cutaway drawing of a Magnox reactor, which gives some visual scale to the monster that is a gas-cooled reactor:

(PDF) http://econtent.unm.edu/cgi-bin/showfile.exe?CISOROOT=/nucen...

(Note this isn't the AGR that Stross depicts, but its predecessor the Magnox). For size reference, the six boilers are 118 feet tall by 18 feet diameter (36m * 5.5m); the building is 170 feet high (52 m).

Reply on news.ycombinator.com