US Nuclear Power Enters the Digital Age

An anonymous reader writes "South Carolina's Oconee Nuclear Station will replace its analog monitoring and operating controls with digital systems, as part of a $2 billion plant upgrade by its owner, Duke Energy. It will become the first nuke plant in the US to use digital controls, and its upgrade may be quickly followed by others. The main driver for the move is cost savings; worries about reliability and hackers have been the reason digital systems haven't been adopted sooner."

This should work out well..

[SuperCharlie]

South Carolina's Oconee Nuclear Station will replace its analog monitoring and operating controls with digital systems

Chinese Military Admits Existence of Cyberwarfare Unit

Wait..

Re:What could possibly go wrong?

[Radworker]

And I suppose your opinion is based on something other than hear-say? Like maybe a little personal experience? Until then I suggest you avoid putting your foot in your mouth. I worked in the industry for 20 years and while I wouldn't paint them as choir boys, I know that the Corporate bean counters aren't the demons you portray them to be.

Re:This is actually scary

[khallow]

I googled around and all I found was some stories about Duke partnering in "clean energy technologies" with a dominant (and probably partly state-owned) Chinese electricity provider. So what is the nature of this relationship with China?

Re:What could possibly go wrong?

[khallow]

Given the arrogant and secretive corporate culture of current nuclear power companies, nothing we'll ever hear about anyway.

Why in the world would a corporate culture, arrogant and secretive or not, want to have anything to do with a bitter, whiny Slashdot drone such as yourself?

Futher, any attempt at cooperation or openness by a nuclear plant operator is seen by the anti-nuke forces as either weakness or some sort of ploy. As a result of this adversarial relationship with a large portion of the population, there's little reason for nuclear operators to volunteer anything beyond what is legally required.

Slashdot fanboys will still love them though.

Here's why I'm a fanboy. Like most of our industrial infrastructure, nuclear plants help build civilization. I don't mind having them compete on even grounds with the other means of producing power, even if nuclear fails hard as a result. But I'm not going to hamstring nuclear power just because it has a corporate culture you don't like.

Re:Ooo! I can solve that one!

[Anonymous Coward]

Not really, it's been shown again and again that if you just drop off enough infected usb keys at an employee parking lot, during a morning or during lunch, that those employees will pick them up and naturally look up what's on those usb keys as soon as they get back in their office.

Hacers not the main problem with all digital I&

[notany]

The biggest problem with digital I&C is the “software common cause failure issue"

Imagine modern nuclear plant with multiple-channel redundancy in instrument and control systems, if one instrument fails, there are others. Same applies to whole cooling systems, if one cooling system fails, there are other completely independent systems that continue to work. Typically redundant systems use instruments from different manufacturers or instruments that are implemented with different technology.

This is not possible for digital systems because they are too costly to implement multiple times. What this means is that redundant digital control systems use same software. If one system fails because of software error, others may follow. This has already happened in German nuclear plant that had new digital system installed. Only the old analog system that was still operational saved the reactor.

This is why Finnish radiation and nuclear safety authority required changes in Areva's plans for the most modern nuclear reactor being build, Olkiluoto 3. They added analog safety requirements. Reactor must be able to shout down even when digital I&C has total failure. Relying for all digital systems compromises redundancy.

More info:

http://www.neimagazine.com/story.asp?storyCode=2053091 [neimagazine.com]

http://www.amazon.com/Digital-Instrumentation-Control-Systems-Nuclear/dp/0309057329 [amazon.com]

Re:Hackers?

[Anonymous Coward]

I can't comment on Points #1, #2, or #4, but I worked in a hospital network for several years and I can tell you that sites were blocked for very good reasons. Like the time we found out 40% of our internet bandwidth was being sucked up by internet radio, ESPN.com, Youtube, and Weatherbug (a few packets every few min is one thing, a few packets every few minutes from 10,000 computers going out the firewall at once for no good reason is something else). As for doctors needing stuff for legitimate reasons? Let me tell you about the Department head that got his team exempted from the internet filters because his team was too important to be second-guessed. We had to get a network tech to go down & muck out all the donkey porn popups every three days. This continued until the female network tech decided that she was sick of knowing what these elite doctors did with their hospital-provided computers & threatened to sue for a Hostile Work environment unless we either A) Re-Blocked the doctors or B) Stopped making the network techs clean up the computer (effectively making it unusable).

Re:Ooo! I can solve that one!

[Raven737]

I looked up how Stuxnet works because it was relevant to my work and company (we use a lot of S7 PLCs on our production network).

The original was now much more than a glorified backdoor. It would install itself but did not contain any directly malicious payload. It would try to connect back to attacker, then the attacker could send and execute any payload they wanted.

It is likely the first payloads where used to identify priorities the attacked system (downloading source code etc). Then a malicious attack payload was specifically created to do the most harm and sent.

It was a glorified backdoor because it could propagate by itself and had the components to detect and connect to, upload and hide code to PLCs.

If it was installed by USB on a PC that was not connected to the internet then it would not have caused any direct harm since it wouldn't have been able to connect to the attacker.

Anyway, of course you can design a variant of Stuxnet that can try to damage any PLC without prior knowledge (contain a malicious payload), but i doubt it would be very effective. Without knowing what a PLC does / is supposed to do, the damage by simply changing values would likely be minimal and be immediately recognized.

Re:What could possibly go wrong?

[Anonymous Coward]

No, not thanks to imports. Thanks to an increase in renewable energy production. The total production of renewable energy grew from 37.8TWh/year in 2000 to 103TWh/year in 2010. That's a factor of 2.7. If you extrapolate that growth to 2022, the increase in renewable energies alone would make nuclear production obsolete (almost steady 150TWh/year over the last 20 years).

Extrapolating over ten years is obviously fishy, but consider that the renewable energy figures include hydro power, which accounted for two thirds of the 2000 renewable energy production and has not increased at all. Photovoltaics have grown from basically 0 to 12TWh/year, wind from 9.5 to 36.5TWh/year, biofuels from 1.6 to 28.4TWh/year (all 2000-2010). As you can see, the growth rates of the energy sources which are responsible for the renewable energy growth in that timeframe are actually much higher than the 270% which includes stagnant sources, so 270% overall renewable energies growth in the next ten years appears to be a rather conservative estimate.