Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Morbius comments. Split from Three Mile Island (survey)

  1. Oct 14, 2004 #1


    User Avatar
    Science Advisor
    Dearly Missed


    The above statements are untrue.

    True, Chernobyl was a graphite moderated design - but that in and of itself
    was not the design defect. One CAN make a safe graphite moderated
    design - like the HTGR [ High Temperature Graphite Reactor ]. For years,
    the Peach Bottom Unit 1 HTGR operated safely in Pennslyvania.

    [ Ironically, it's better to operate the graphite reactor at high
    temperature. The higher temperature anneals the build up of
    "Wigner Energy" - so that it doesn't start a graphite fire as it
    did at the Windscale production reactor in Great Britain in 1957



    Additionally, the Chernobyl RBMK reactor was NOT designed in Hanford.

    One basic problem with the Chernobyl RBMK reactor design is that it is
    a "scaled up" version of a weapons fuel production reactor like the ones
    at Hanford. The Russians built their versions of the Hanford reactors.

    However, in designing the Chernobyl RBMK reactor - the Russians basically
    built a 2 X 2 X 2 stack of their smaller production reactor. However,
    they did not redo the nuclear design of the reactor.

    When you stack reactors like that - you reduce neutron leakage. The
    neutrons that would have leaked out of the reactor - leak into the one
    next to it, and vice-versa. Therefore, the combined stack leaks less
    than the smaller reactors.

    Because of the reduced leakage, the RBMK is "over moderated"- which
    makes it unstable. All USA reactors have been "under moderated" and
    are required by law to be such - as this makes the reactor stable.

    Additionally, the Russians did some rather poor design of the control
    rods - like having fissile fuel "followers".

    The Chernobyl accident was triggered by an experiment the operators
    were performing on the reactor. Because of a delay, this experiment
    was performed in the middle of the post-shutdown "Xenon transient"
    that all reactors go through. The reactor in that state was highly
    unstable, and the operators could only get it to run by bypassing many
    safety interlocks. The result is history.

    Dr. Gregory Greenman
    Physicist LLNL
    Last edited: Oct 14, 2004
  2. jcsd
  3. Oct 14, 2004 #2


    User Avatar
    Science Advisor
    Dearly Missed


    This part of your post is also incorrect.

    The hydrogen in the bubble was produced mainly by zirconium oxidation.

    The fuel elements are clad in a metal coating of zirconium. If you heat
    the zirconium to a temperature that is high enough in the presence of
    water, it has a greater affinity for the oxygen in the water than the
    hydrogen does.

    The oxygen in the water reacts with the zirconium to form zirconium oxide
    and free hydrogen. Hence there are no hydroxyl ions left in the water.

    [ There's no free oxygen either - only free hydrogen - which is why it
    can't explode.]

    A very small part of the hydrogen was formed by radiolytic decomposition.
    In this case, yes - there would be free hydroxl ions. However, as you
    point out, the inverse reaction is also permitted and the hydrogen
    and the hydroxyl radicals recombine - thus limiting the amount of
    free hydrogen and hydroxyl radicals produced this way to a very small

    Dr. Gregory Greenman
    Physicist LLNL
    Last edited: Oct 14, 2004
  4. Oct 14, 2004 #3


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    I split these comments from the survey, Morbius. Please keep any discussion of the posts in idonthak's survey here.


    -Engineering Mentor

  5. Oct 14, 2004 #4


    User Avatar
    Science Advisor
    Dearly Missed


    Sorry about the Off Topic posts.

    Dr. Gregory Greenman
    Physicist LLNL
  6. Oct 14, 2004 #5


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    No problem.

    Ordinarily, they would have been entirely appropriate. It's just that he was posting a survey, and I didn't want a side conversation to detract from that.

    Welcome to the forums, BTW!
  7. Oct 20, 2004 #6
    Is this an intended safeguard? Very intresting!
    One more thing, why isn't there any free oxygen? Does all of it combine with the zirconium?
  8. Oct 20, 2004 #7


    User Avatar
    Science Advisor
    Dearly Missed


    It's not a safeguard - it's just chemistry. In fact, the design intent is
    to minimize the amount of zirconium oxidation.

    You've got it correct - the only way hydrogen is made in this reaction is
    that the water molecule H20 is split into H2 and O, and the O combines
    with the Zirconium.

    So there's no free oxygen - the hydrogen that is produced is what is left
    after the hot Zirconium grabs the Oxygen atom out of the water molecule.

    That's why the hydrogen bubble in the reactor can't explode. Hydrogen
    gas is not explosive. A gas that is a mixture of hydrogen and oxygen
    CAN explode.

    Since there was no free oxygen in the reactor - there was no way that
    the hydrogen could explode. [ It took a while for people to realise this ]

    Since the relief valve was stuck open - allowing coolant water to gush
    into the containment building - some of the hydrogen escaped the
    reactor that way - and out into the containment building where there
    was plenty of air with free oxygen. There the hydrogen DID EXPLODE!

    However, this explosion - outside of the reactor - could not harm the
    reactor itself. It also didn't harm the containment building since the
    pressure from the hydrogen explosion was far less than what the
    containment building was designed to withstand.

    So, in the end - the hydrogen bubble was a non-issue as far as the safety
    of the public. The public was needlessly scared by the hydrogen bubble

    Dr. Gregory Greenman
    Physicist LLNL
  9. Nov 14, 2004 #8
    Can you go into some technical depth here (or in a message to me) possibly listing some equations and material properties?
  10. Nov 15, 2004 #9


    User Avatar
    Science Advisor
    Dearly Missed


    I can give you an explanation right here.

    There is an optimal amount of moderator that maximizes the reactivity
    of a critical assembly. If you plot the reactivity of the system vs. the
    amount of moderator - you will get a "hump" shaped curve. Check any text
    in nuclear reactor physics.

    All USA licensed reactors are required to be under-moderated - that is to
    have an amount of moderator that is less that that which gives maximal
    reactivity. Hence, they are on the uphill side of the reactivity curve to
    the left of the maximum.

    When a reactor loses coolant - it also loses moderator. If the reactor
    is under-moderated - the decrease in moderator with the loss of coolant
    shifts the point on the curve to the left - and since it's on the "uphill"
    side to the left of the peak - the shift to the left means the reactivity
    goes down. The reactor is self-stabilizing when it is under-moderated.

    Contrast this to Chernobyl, which was over-moderated. The operating
    point for Chernobyl was to the right of the peak on the "downhill" side
    of the curve. When Chernobyl lost coolant, and hence moderator - the
    the operating point shifted left which takes it UP in reactivity!!!

    Chernobyl had too much moderator - so when the accident removed
    some of the moderator [ the coolant water ] - the reactivity increased!

    That, addition to the fact that the operators were operating the reactor
    when it was Xenon poisoned following a reduction of power - meant the
    RBMK reactor was in an unstable state - and the result is history.

    Dr. Gregory Greenman
  11. Nov 15, 2004 #10
    Does that mean there is a hump in the cross section vs neutron energy curve?

    That looks like a good deal.

    What kind of curves/ equations do you have on the Xenon performance characteristic? Say reactivity vs neutron flux?

    I have the Physics Today issue on Chernobyl somewhere buried in a box.
  12. Nov 16, 2004 #11


    User Avatar
    Science Advisor
    Dearly Missed

    NO. In general, the cross-section is a very complicated function of energy.

    Reactivity is not a strong function of neutron flux.

    How criticality or sub-criticality of a nuclear assembly is NOT dependent
    on the flux - but solely on the materials and the geometry of the reactor.

    The assembly can be critical - even if the neutron flux is zero. Of course,
    nothing happens until you get some neutrons.

    Reactivity is not totally independent of neutron flux - but to first order,
    it is dependent mostly on geometry and materials.

    As far as Xenon poisoning is concerned. One of the most plentiful fission
    products is Iodine-135. Iodine-135 decays to Xenon-135. Now Xe-135 is
    THE world champion neutron absorber!! It is destroyed in the reactor in
    two ways - it absorbs a neutron, or it decays radioactively.

    When the reactor is at power - there is an equilibrium level each of I-135
    and Xe-135. When the reactor is shutdown, or the power is lowered, the
    source of I-135 is stopped or reduced, and the destruction of Xe-135 by
    neutron absorption is stopped or reduced.

    However, immediately after shutdown - the level of I-135 is the same as
    the equilibrium level - just because you shutdown the reactor doesn't
    mean the I-135 goes away.

    Since the I-135 level is still equal to the equilibrium level - the source of
    Xe-135 doesn't change. But one of the destruction mechanisms is stopped.

    Since there was an equilibrium between the source rate of Xe-135 and
    the destruction rate of Xe-135; and the source rate is effectively
    unchanged, while the destruction rate is reduced - the net effect is an
    increase in Xe-135 levels - which poison the reactor core.

    Eventually, the I-135 decays away, so the source is reduced, and without
    a source - the Xe-135 decays away. However, that process takes the
    better part of a day.

    At Chernobyl, the operators lowered the power to run an experiment on
    the reactor. The load controller in Kiev called and asked that the plant
    stay online at the reduced power level. After 12 hours, the load
    controller released the plant from the electrical grid.

    At 12 hours, the Chernobyl RBMK reactor was in the middle of the
    Xenon transient - it was experiencing maximal Xenon levels.

    The operators were having a difficult time maintaining criticality - they
    pulled all the control rods and bypassed safety systems that were trying
    to shutdown the reactor because it was in an unstable condition!

    Unfortunately, the operators triumphed over the safety systems and
    proceeded to run the experiment! The RBMK is marginally stable under
    normal conditions - but in the state the operators put the reactor in
    while in the midst of a Xenon transient - it's hardly any wonder that they
    couldn't control it.

    Dr. Gregory Greenman
  13. Nov 16, 2004 #12


    User Avatar
    Staff Emeritus
    Science Advisor

    Modern light water reactor (LWR) fuel and core configurations are designed for negative moderator temperature coefficients, i.e. they are designed such that if the coolant (moderator) temperature increases , and the moderator density decreases (beyond the normal range), the reactivity decreases, so there is a 'natural' control inherent in the reactor design.

    All commercial nuclear power reactors in the US are LWRs, which use water as both moderator and coolant. In Boiling Water Reactors (BWRs) the water is boiled directly in the core, and the steam is passed directly to the turbine train to produce power. Reactivity is controlled directly with the control blades (using boron in B4C or Hf enshrouded in stainless steel) located between the fuel assemblies and use of burnable poisons (gadolinium in form of oxide) in some of the fuel rods. During the cycle, the control rods are withdrawn gradually as fission products (which also function as neutron poisons) accumulate in the UO2 fuel fuel.

    In Pressurized Water Reactor, the water is pressurized to limit the boiling to some amount of so-called nucleate boiling (heat is passed from primary (reactor cooling) system through a steam generator and steam is produced in a physically separate circuit). The control rods are withdrawn above the core during operation, so reactivity is controlled by soluble boron (in the form of boric acid, with an appropriate buffering agent like LiOH) in the coolant as well as burnable poisons in the fuel (oxides of gadolinium or erbium, or coating of ZrB2 on the surface of the fuel pellets).

    In either case, as the coolant density decreases below normal range, the moderation decreases and the reactors go subcritical. If there is some significant perturbation in the reactor operation, the control rods are inserted automatically and the reactor is shutdown.

    RBMK's are graphite-moderated reactors using water as coolant. It turns out that the water in such a system behaves as a neutron absorber (light water - with H and O has a reasonably good neutron capture cross-section) and as the water heats and density decreases the reactivity increases, which is the opposite of the way the water moderator in an LWR behaves.

    Not remembering the details, I presume the Xe in the Chernobyl 4 reactor was decreasing and the reactivity began to increase. Also as the neutron flux increased, it would start to 'burnout' or consume the Xe-135. As the heat increased, the liquid water turned to steam, which further increased the reactivity - the power continued to increase. Since the control rods (safety systems) had been over-ridden, the staff could not respond fast enough to reinsert them - they probably had a matter of seconds once they realized (if they did) what was happening and the rest is history.

    For some background on RBMK, see - http://www.world-nuclear.org/info/inf31.htm [Broken] - which discusses somewhat the concept of "Positive void coefficient" in the coolant/moderator.

    As for TMI-2, it was in its third month of commercial operation when the problem occurred. Lack of experience contributed to the problem. As a result, all reactor operators receive constant training, testing and assessment. Training includes very detailed simulators that simulate normal and abnormal reactor operating conditions. If the operators fail the simulator or training - they are not allowed to operate the reactor.
    Last edited by a moderator: May 1, 2017
  14. Nov 16, 2004 #13
    Somehow I am not getting to the relationship that I want. How does the excess Xe-135 cause instability? Instability means that as N goes up then K goes up. Pulling rods raises K. Stability means that as N goes up then K goes down. How fast did the power rise at Chernobyl?
  15. Nov 17, 2004 #14


    User Avatar
    Science Advisor
    Dearly Missed


    Not quite - the limit is not "nuclear boiling" [ there's no such thing].
    I believe the term you are searching for is "nucleate boiling".

    Specifically, the limit is controlled by concern for "departure from
    nucleate boiling". The type of boiling that you see in a saucepan on the
    kitchen stove - where little bubbles form on the bottom of the pan is
    called "nucleate boiling". There's a very limited amount of that type
    of boiling going on inside a PWR.

    What one wants to prevent is "departure from nucleate boiling" which is
    when the temperature of the heated surface exceeds the "Leidenfrost"
    temperature. If the temperature of the surface exceeds the
    Leidenfrost temperature - the water won't "wet" the surface.

    Next time you fix pancakes - try a little experiment with the griddle
    before you cook the pancakes. As the griddle heats up - drop a couple
    drops of water on the griddle. You will see the drop spread out on the
    surface of the griddle and very quickly evaporate.

    However, if you get the griddle hot enough - the drop will "dance around"
    on the griddle surface for several seconds - much longer than the life
    of the drop that spread out on the griddle surface.

    When the drop dances around like that - the griddle has exceeded the
    Leidenfrost temperature. The griddle is so hot that where the drop
    touches the griddle it flashes a small part of the drop into steam. This
    steam forms a layer that insulates the water drop from futher heating,
    and allows it to dance around. As the steam layer diffuses away, and
    the drop again can momentarily contact the griddle surface - there will
    be more steam made - and again the drop will be insulated. Eventually,
    all the drop will have been turned to steam - but it takes longer than if
    the water wet the surface - as is the case at lower temperatures.

    One wants to prevent this "vapor blanketing" and insulating effect of
    steam production. When there's an insulating steam layer - the heat
    transfer from the heated surface goes down - and the temperature of
    the heated surface goes up - because the coolant is being less effective
    at doing its job.

    There is a limit called the DNBR - Departure from Nucleate Boiling
    Ratio. It is the ratio of the heat flux at which one would get DNB -
    Departure from Nucleate Boiling - i.e. exceeding the Leidenfrost temp.
    to the heat flux in the reactor. This value is specified in the reactor's
    "Tech Specs" - the limits of which are conditions of the operating license.

    If the DNBR, was given as 0.85 - then the power of the reactor is limited
    so that the heat flux stays more than 15% away from DNB conditions.

    This is only the case if the reactor is "over-moderated" - like the RBMKs
    are. One can build a graphite moderated reactor in which there is not
    enough graphite to do all the moderation required. That means that
    some of the moderation has to be done by the coolant water. If the
    coolant water is providing needed moderation - then you get a
    negative coolant temperature coefficient - just like in an LWR!

    Not quite. At about 12 hours since the power was lowered the Xenon
    levels were not yet decreasing. The problem was that the core was so
    poisoned with Xenon - the operators fully withdrew the control rods.

    The RBMK also had a bad control rod design - one in which the rod had
    an active fuel "follower". So during the first part of a rod insertion -
    the rod was actually adding more fuel to the core - before it added
    the neutron absorber.

    During my graduate studies at M.I.T., I also attended a seminar by
    Professor Kemeny - who led the commission that investigated the TMI
    accident. At one point, while visiting the TMI control room after the
    accident - Kemeny asked the operators for a "steam table" - that's a
    book that gives the Equation of State of Water - that is for each value
    of temperature and pressure - it tells you whether water is liquid, vapor,
    or boiling. Kemeny said it took the TMI operators about 15 minutes to
    find a steam table!

    The operators at TMI didn't know how far away from boiling conditions
    they were. Kemeny said anyone who read the accident chronology in the
    newspaper understood the problem.

    I knew EXACTLY what he meant. I remember reading the accident
    chronology in the newspaper. At one point, the reactor operators
    reported that they had "stabilized" the reactor at such and such a
    temperature and such and such a pressure.

    I wondered how far they were from boiling. So I reached up to get my
    copy of Keenan and Keyes steam tables which was sitting on my filing
    cabinet - and looked up the Equation of State for Water at the
    conditions specified in the newspaper.

    Those conditions were right ON the "saturation line" or "boiling curve".
    I said to myself - "They didn't "stabilize" the reactor - they're BOILING!"

    The reason the temperature and pressure in the TMI reactor stabilized
    was that the water was boiling - and the operators thought everything
    was fine - that they had solved the problem - when the worst possible
    thing was currently happening!!!

    Dr. Gregory Greenman
    Last edited by a moderator: May 1, 2017
  16. Nov 17, 2004 #15


    User Avatar
    Staff Emeritus
    Science Advisor

    Morbius, Thanks for the correction - I did mean 'nucleate' boiling as you described.

    With respect to Departure from Nucleate Boiling Ratio (DNBR) -

    [tex]DNBR = \frac{{\large{q}\prime\prime}_{DNB}}{{\large{q}\prime\prime}_w}[/tex]

    where [tex]{\large{q}\prime\prime}_{DNB}[/tex] is the heat flux that causes DNB under local temperature and pressure conditions, or perhaps more accurately is the heat flux predicted to cause DNB under the local conditions according to some correlation (usually based on ex-core experiments)

    and [tex]{\large{q}\prime\prime}_w[/tex] is the local wall heat flux under the operating conditions.

    Of course, there are more practical considerations, such as corrosion of the Zr-alloy cladding. The heat flux is limited in order to limit the local cladding temperature in order to prevent excessive cladding corrosion and crud (metal oxide corrosion products from the primary cooling circuit surfaces, primarily steam generator tubing). Some heavy crud deposits have been associated with reactivity anomalies in some reactors - but this is now controlled (avoided) by limiting coolant temperature and better controls on water chemistry.

    In operating PWRs, typical DNBR's have been on the order of 1.26 or 1.3, a margin of 0.26-0.3 to DNB. The relatively high margin was due to uncertainties in the DNB (of CHF) correlation and the uncertainties in the core monitoring systems. With improved models and monitoring systems, some plants can reduce the DNBR to about 1.17 (IIRC), but it is always greater than one by definition.

    There are some European PWRs that operate with considerable levels of nucleate boiling, but plants in the US operate with much less.
  17. Nov 18, 2004 #16


    User Avatar
    Science Advisor
    Dearly Missed


    Exactly - the limit is for the most limiting local value of DNBR.

    Thank you for refreshing my memory on typical values of DNBR.
    [ I may have also inverted the ratio ].

    It's been a long time since I've had to contend with those considerations.

    Dr. Gregory Greenman
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook