Welcome to European Tribune. It's gone a bit quiet around here these days, but it's still going.
Due to the voluminous design of the reactor core it is impossible to add a full containment to the reactor. The RBMK is basically the only reactor type lacking a full containment building.

I think you understate seriously the danger of RBMKs. Sure, the lack of containment is dreadful but RBMKs are much worse than that.

They have a positive void coefficient, meaning that if a RBMK loses some of its cooling water, the core reactivity goes up and boils off more water, which boosts the reactivity even more and so on and so forth all the way to steam explosion and meltdown. A design with a positive void coefficient can be ok if there is another strong, intrinsic negative feedback designed in the reactor, such as free expansion of the core assembly in metal-cooled FNRs, loss of moderation in PWRs and BWRs or Doppler broadening in high-temperature reactors (PBRs for instance). But without such a counter-acting effect built in the core physics, positive void coefficient is the worst conceivable sin of nuclear reactor design.

The RBMK design is evil, absolutely evil (*). Not only, it cannot receive a containment shell because its ludicrous height but it is the only design operating today that can actually blow up. There should be a containment shell and it should be one hell of a massive bunker, something like 20 or 30 meters thick of reinforced concrete and that ain't gonna happen.

All RBMKs must be closed, no if, no but. Europe should be pounding fists and feet on Lithuania if they back off.

(*) At least, the design of the control rods has been fixed, or at least I think it has been fixed ...
by Francois in Paris on Sat Apr 29th, 2006 at 11:09:20 PM EST
My understanding of the RBMK principal design motivation, is that by having moderate pressures inside (and hence letting whe water boil), you can arrange for systems for hatches to pass fuel rods in and out of the core during operation, thus permitting zero-downtime refuelling.

I don't know the procedural details of this refuelling, but it certainly makes for an intrinsic weakness of the reactor vessel: PWR are basically ripped apart and soldered again when refuelled, and there is no way out of the vessel (except melting it...) when it's sealed and coolant loops are closed (and the primaries don't even extend out of the vessel for integrated PWR.

By contrast, a RBMK has a some sort of passageway with moving shutter for the fuel rods, so the primary containment is weaker. This indeed makes it evil. As for the negative void coefficient, I don't know it this is bound to happen in any runtime-refuelable water moderated design, or only an added design fault of this soviet implementation of the concept. Note that doppler broadening is already effective in Westinghouse-type PWR (e.g. french plants), it's a passive feedback loop that's alone 100% efficient in maintaining core temperature withing a few degrees of the design goal during "cruise" operation.

Anyway, I think the PBR's have a decisive edge these days when you need runtime refueling.

by Pierre on Sun Apr 30th, 2006 at 07:08:14 AM EST
[ Parent ]
Update on this: Areva still has a boiling-water reactor in its offering, though they're not marketing it as strongly as the EPR. It is a remain of the Siemens business, from before the merger with Framatome.

Since it's a western, german, recent design, I cannot imagine that those german engineers would dare offering a reactor with an instable void coefficient. (Although I have not found any statement on this on the Areva website).

by Pierre on Sun Apr 30th, 2006 at 04:09:00 PM EST
[ Parent ]
Well, yes but no.

Boiling water is not by itself an issue as demonstrated by the very satisfactory operating record of BWR reactors in the US and Western Europe. The idea of using the primary coolant to directly spin the alternators is a bit freaky at first look but, in practice it works well, with the added benefit that it induces very rigorous safety practices.

The issue is that the RBMK design is a dual-use reactor. Producing power is one of its goals but the most important goal is to produce weapon-grade plutonium. Even if it was outwardly a civilian operation, Chernobyl is, like every real nuclear disaster, deeply related to military use of nuclear energy.

Every single flawed design choices made in RBMK derives from that breeding capacity. At the top of the list comes the choice of graphite moderation with ordinary water coolant. With this design, the ordinary water that flows along the fuel assemblies and vaporizes plays the role of a neutron absorber, not that of a moderator as in PWRs or BWRs. It means that to compensate neutron loss in the coolant, the design must have higher reactivity than it would need to maintain criticality if the reactor was dry. So, if too much coolant vaporizes (excess power), less neutrons are absorbed by that missing water and those excess neutrons reach the graphite moderator, get moderated and reach fuel assemblies to create more reactions. The power of the reactor goes up and so on.

I believe (not 100% sure) that under normal operation of a RBMK, the stabilizing effect to excess power is that if there is a localized hot spot, the pressure tubes around the hot spot generate too much steam (bad) but also make the pressure rises in the whole reactor, bringing more water back in dense liquid phase over the whole core so more neutrons are absorbed and overall power goes down (good).

But, in the case of a large, sudden breach in a pressure tube, the reactor goes dry with excess reactivity and a reaction runaway is guaranteed. And at Chernobyl, the whole pressure system ruptured at once with the power spike. The same effect, that ordinary water is not a moderator but an absorber in this type of reactor, also explains the role of the graphite tips on the control rods in the lead up to the power spike.

Mixing graphite and water is simply evil. Note that the US military program used uncomfortably similar systems for weapon-grade plutonium production at the Hanford site. All reactors were closed between 1964 and 1971 except for one, Hanford-N, which, interestingly enough was also a dual-use plutonium/power reactor. It was closed in January 1987...

I did not mentioned Doppler broadening as a safety feature in PWRs and BWRs because, as you mentioned, it plays a role in driving the reactor, but it plays no significant role in a catastrophic LOCA situation.

The big last line of defense when a PWR or BWR suffers a massive loss of coolant (like at Three Miles Island) is that the coolant is also the moderator. If the cooling water evaporates out of the core, the neutrons are not moderated anymore. They leak outside of the active core area without having the time to reach levels of energy low enough to sustain the fission reaction while in the core. The reaction stops and the only remaining power output in the core is the residual decay heat from short-lived fission products, about 10% of the operating power. TMI was bad but did not come anywhere close to Chernobyl in good part for this reason. The core slowly self-destroyed but could not really explode.

This mechanism also plays a role in the normal operation of BWRs. The reactor are protected against vaporization instability because too much steam in the core reduces moderation and, as a consequence, it reduces the reactivity, hence the power output. Nice negative feedback.

PS: You like PBRs ? Personally, I'm not convinced ...
by Francois in Paris on Mon May 1st, 2006 at 02:06:32 AM EST
[ Parent ]
Indeed, having a hidden agenda (like generating plutonium or other dual-purpose, incompatible uses) is not a good principle when building a utility-size reactor (not that the British didn't do very well with windscale either, talking about non sense like air-cooled graphite moderated, though it was a small and openly military-only installation).

Note that the german BWR has a separate secondary coolant loop to drive the generators. I believe US reactors would have that too.

PBR's seem convincing to me yes, for a number of reasons, like:
  • passively safe

  • modular plant with small power increment, fit small grids, or even non-grid use like chemical process heat generation

  • compact vessel, suits a mass-manufacture-and-lease-then-take-away model, including for shipping to developping countries

  • ability to adapt power output quickly, to match peak demand, or simply to complement irregular renewables just like hydro or gas does

  • plus, there was actually a running implentation for years at Julich, which actually makes it just as proven as the EPR

  • Of course, there is the story of the jammed pebble at Julich, plus the (often untold) detail of the pebble cooling facility at the output of the vessel, before sifting pebbles to loop back into the vessel or to dispose of, which is a big security/safety weakness, plus it spoils somewhat the compacity thing.

    by Pierre on Mon May 1st, 2006 at 03:52:30 AM EST
    [ Parent ]
    Thanks guys for all the detailed technical input.

    In the long run, we're all dead. John Maynard Keynes
    by Jerome a Paris (etg@eurotrib.com) on Mon May 1st, 2006 at 05:24:16 AM EST
    [ Parent ]
    Are you sarcastic for the nerds, Jerôme :-?
    BTW I'd still like to read your opinion on my post regarding the impact of GTL on the gas and oil markets, here...

    by Pierre on Mon May 1st, 2006 at 06:42:27 AM EST
    [ Parent ]
    I don't think it's sarcasm: Jerome is a reality-based banker.

    A society committed to the notion that government is always bad will have bad government. And it doesn't have to be that way. — Paul Krugman
    by Migeru (migeru at eurotrib dot com) on Mon May 1st, 2006 at 06:48:30 AM EST
    [ Parent ]
    No, it was honest appreciation!
    As to GTL, it's a topic I know too much about and decided I could not do a short reply, and did not then have time for a longer one...

    I actually wrote a diary on this a year ago: How to make gasoline from gas

    In the long run, we're all dead. John Maynard Keynes

    by Jerome a Paris (etg@eurotrib.com) on Mon May 1st, 2006 at 07:48:11 AM EST
    [ Parent ]
    Mixing military and civilian reactors is never a good thing. Our only somewhat nasty accident happened at the Ågesta heavy water reactor, which was probably a joint effort between the military and the civilians.

    Peak oil is not an energy crisis. It is a liquid fuel crisis.
    by Starvid on Mon May 1st, 2006 at 07:45:17 AM EST
    [ Parent ]

    I confirm that in a BWR, the steam goes directly from the core to the turbine, including on the latest Siemens design. Product brief here. Very nice for conversion efficiency and the pressure stress on the core vessel is lower than for a PWR (operation at 75 bar for SWR-1000 vs. 155 bars for EPR). The issue is of course that the turbine is slightly activated by the primary steam and that a breach in a fuel cladding is much more annoying than with a PWR. The fuel assembly is also more complex and the control rods cannot be placed in the fuel channels (risk of being pushed out by the steam) so it has to be installed on the sides of the assembly cans. BWR is a different trade-off than PWR. Gain some, lose some, the eternal story of engineering.

    I'm pretty skeptical of PBRs.

    I find the idea of making the fuel hard to reprocess very offensive. Wasting all that good plutonium and that excellent fertile U238, what a shame. Just to think of all that energy sends shivers down my spine :)

    More seriously, the graphite pebble casing makes the fuel very bulky and hard to dispose. Burial of anything but ultimate waste is not a valid option in my view and even then, it should be reversible (I'm not very happy about Finland). Separation to recover the spent fuel safely is a no-no as it would creates a lot of activated graphite waste in dust form (C14 is biologically nasty). May be PBMR promoters think it's a smart idea but IMHO, it's about the worst you can do for waste management.

    Same for the non-proliferation claim. It looks pretty bogus. The large number of elements makes inventory control difficult and theft easy. A proliferator can purloin pebbles a few at a time, and by crushing and milling, recover the spent fuel (and create lot of activated dust, as mentioned above, but a proliferator would not mind). If proliferation is really a concern, fully canned reactors seem a better option such as Livermore's SSTAR concept. See that for more details.

    I'm also non-plused by the safety claims. The passive safety by Doppler broadening against thermal runaways is fine and proven but ...

    - The PBMR design has no containment at all. Whaa ?!? Big conventional reactors are of course fully containted but they are not modular and not suitable for operation in poor countries. For a more honest comparison, let's take the SSTAR design. The SSTAR is not fully contained neither but it is at least buried, all vessel breaches are contained within a heavy concrete casing, the vessel is mechanically much, much sturdier than a PBMR tank and the core is isolated in the middle of a large mass of lead. The only area exposed to ambient air is the wall of the bottom of the core vessel for emergency cooling. It's very thick and built as a single piece forging with no breach. You gotta try really hard to expose the core.

    - I haven't seen anything convincing about cold air inrush in case of a rupture at the vessel at the inlet of the coolant compressor. The pebbles would be exposed to a very brutal thermal stress so how about cracking a pebble, exposing a highly ragged and dispersed section of carbon to air and have a nice little bonfire? I know that pyrolytic carbon doesn't catch fire just like that but I would really like to see serious experiments with pebbles that have gone through multiple loading-unloading cycles with all the accompanying mechanical stress.

    My preference for future design rather go to metal-cooled fast reactors and, longer term, to molten salts reactors with continuous fuel control.

    And with that comment, promised, I'm done with nuclear geekiness for the week :)
    by Francois in Paris on Mon May 1st, 2006 at 01:54:16 PM EST
    [ Parent ]
    Will do. :)

    Thanks for all the comments on my first diary ever! :)

    Peak oil is not an energy crisis. It is a liquid fuel crisis.

    by Starvid on Mon May 1st, 2006 at 07:30:19 AM EST
    [ Parent ]
    Sorry wrong comment on this post, reposting it.

    Peak oil is not an energy crisis. It is a liquid fuel crisis.
    by Starvid on Mon May 1st, 2006 at 07:32:01 AM EST
    [ Parent ]
    I didn't mention all the other nastiness of the RBMK because I figured that at the twentieth anniversary of Chernobyl, people would by themselves understand RBMK's are a very bad idea.

    I agree they are all pure evil and should be closed as fast as possible. Still, they (especially Ignalina) are a lot safer then they used to be.

    Peak oil is not an energy crisis. It is a liquid fuel crisis.

    by Starvid on Mon May 1st, 2006 at 07:37:31 AM EST
    [ Parent ]


    Occasional Series