Mother Nature Does Nuclear Power

wjwlsn writes "Back in the day (2 billion years ago), even before the time of iron men and wooden reactors, Mother Nature had mastered nuclear power. She built a passively safe system at Oklo that had fully automatic control and built-in waste containment, and operated it safely for about 150 million years. Now researchers at Washington University in St. Louis have deduced the operational characteristics by examining the isotopic composition of xenon contained in rock samples taken from the reactor site. More details at Eurekalert."

Comments

Not just 2 billion years ago.

[noselasd]

Even today mother nature does nuclear power

OPOD Link

[Trikenstein]

APOD: 2002 October 16 - Oklo: Ancient African Nuclear Reactors

Re:Time spans

[Anonymous Coward]

Faulty argument, but conclusion is true. This article states that the US electricity is 2% by oil and this article states that 2/3 of all oil is used in cars. We can assume that the other 1/3 is for electricity (with trivial amounts going to other petroleum products). We can also note that the oil percentage has gone up greatly since California had its energy crisis (and decided to make tons of oil fired plants). You can do the math, but your conclusion appears to be true. If we lost the 20% of power being made by nuclear, our dependance on foreign oil might be very scary. I've seen manufacturing companies go out of business because of a couple of percent rise in energy prices. If energy prices fluctuated by 50% (like they do for gas at the pump), manufacturing companies would have a very hard time.

Re:This is news?

[PaSTE]

The sun is a fusion reactor, which is not remarkable by any means--if you put enough light stuff in a tight space, gravity will crush it into a fusion reactor without any sort of quirks or anomolies. What makes this news is that nature had created a fission reactor--something that doesn't just happen if you have a lot of heavy stuff in a tight space. You need enough of relatively uncommon isotopes of Uranium, something created in very, very tiny amounts in supernovae, with enough neutron inhibitor mixed in to prevent a melt-down, but not so much that is prevents the reaction from happening in the first place. Quite news-worthy, indeed.

cool, but not a practcal method

[Mr. Slippery]

passively safe system at Oklo that had fully automatic control and built-in waste containment, and operated it safely for about 150 million years

And with a 30 minute reaction cycle followed by a 150 minute dormant period, in a manner that I would guess is almost useless for power generation.

Re:cool, but not a practcal method

[M1FCJ]

30 minute reaction cycle is a pretty long time for reactions. Nuclear reactions happen in a fraction of a second. Recent tokamak reactors have been operated for around a fraction of a second to a second. Fission reactors are easier to manage but still 30 minutes of reaction is pretty substantial. If it were just a thermal expansion that stopped a reaction (google for the first man to die because if direct irradiation of a nuclear reaction) it would have taked less than a second. On the other hand, the reaction itself is quite slow, the article suggests around 100 kilowatts which is only as good as 100 typical electrical kettles.

If this reactor had more moderator, flowing in like a river does, it would have worked for longer. 30 minutes to boil all of the water around you to generate a geiser is quite spectacular. Water boils away, gets converted to steam, reaction stops because of the loss of moderation, some other water flow in, cools the rocks and everything starts again. Neat.

Many of the reactor designs use this negative reactivity coefficient to stop the reactor running away. This is usually used with the pressurised water reactors where the coolant (and the moderator which is water (heavy or normal depending on the reactor design) flows through the elements with a forced circulation. If you lose the circulation, water boils, reducing the amount of moderator, hence slowing the reaction. If you loose enough water, the reaction will stop. (although it is a little bit more complex, depending on the reaction design, you also lose the ability to cool down the elements (air is much worse heat conductor compared to water, even a boling one).

Reactor design is not simple, there are many things to think about, how to moderate, how to cool down, how not to overheat (this is critical because the claddings around the elements usually get weaker when heated and crack. Once cracked, you cannot stop contaminating the water used for the reactor). Anyway, reactor design is one of the most beautiful engineering challenges I can think of. If only I could work on it even more, it was fun, pure engineering, even in undergrad level, it was a joy to learn. If only I could work on it for longer. When I graduated from my university, I had to meet real-life situations, no one wants a good engineering solution if it is marked "Nukular". :-(

Re:cool, but not a practcal method

[turgid]

Recent tokamak reactors have been operated for around a fraction of a second to a second.

When I visited JET back in 2001 they said they were achieving sustained reactions over several tens of seconds (~30) before the plasma became unstable.

Fission reactors are easier to manage but still 30 minutes of reaction is pretty substantial.

Well, my old powerstation used to manage several months of continuous fission reactions on each reactor, before thunderstorms or welding operations or rod-drops would cause the reactors to come off. In theory, a reactor could be run continuously for 2 years i.e. between statutory (legal) biennial outages. These were reactors designed in the late 1950s.

Reactor design is not simple, there are many things to think about, how to moderate, how to cool down, how not to overheat (this is critical because the claddings around the elements usually get weaker when heated and crack. Once cracked, you cannot stop contaminating the water used for the reactor).

Here in the UK most of our reactors are gas-cooled (using carbon dioxide). We have one commercial PWR in Suffolk (Sizewell B). The Magnoxes were positive-feedback systems and could, in theory, overheat, but in practice the passive safety systems prevented this. The AGRs avoid this problem (caused by plutonium resonance with the thermal neutrons and graphite moderator) by holding the graphite temperature steady, by providing the graphite with it's own cooling loop (actually the first stage of core cooling, the gas then gets passed over the fuel). In effect the cold gas coming in cools the moderator, picking up some heat (being pre-heated) and then cooling the fuel, up to about 650 degrees C IIRC.

This all relies on active feedback systems as it is a chaotic system (in conjunction with the boilers).

If an AGR looses forced cooling, it's quite dangerous, as there is a maximum period of time in which you must get the automatic system back up and running. Otherwise you risk ruining your boilers. The "superheat" part of the boilers must under no circumstances get wet or else they are knackered forever, and your powerstation is useless. (AGRs and the two concrete pressure vessel Magnoxes, Oldbury and Wylfa, have "once-through boilers" which are a unique British design developed specifically for nuclear reactors and used nowhere else in the world).

AGRs are better than PWRs in another respect and that is the reactor pressure vessel is too strong to ever develop a significant breach that would result in a depressurisation and catastrophic release of radioactive substances.

Unfortunately, Margaret Thatcher chose a PWR for Sizewell B to improve Anglo-American relations. PWRs do not have concrete pressure vessels and are more "dosey" that AGRs (and the two concrete Magnoxes). They od have a sealed containment building, whic saved the day at Three Mile Island, but this is not required in an AGR or PBMR since the pressure vessel is much stronger and the failure modes are different. AGRs can not melt their fuel even with no forced convection, as long as you keep water in the boilers.

Re:Time spans

[M1FCJ]

Yep, true. Apparently Oklo genereated around 100 Kilowatt (thremal). A typical nuclear reactor usually generates around 2000 NWatt thermal and 1200 Megawatt electrical.