Friday, March 18, 2011

A technical viewpoint of the Fukushima Daiichi Reactor failure

For the last couple nights, I have spent a lot of time trying to sort out this whole nuclear 'disaster' over in Japan from an engineers standpoint, that is, without all the media distortion. So this will be part one of a multiple part series. All of the material has either been gathered via my own personal research or from first hand resources: primarily my Grandfather who is a nuclear engineer and knows way more about these things then the media will ever. He also worked at the Power Burst Facility which sought to test what would happen to a reactor in a situation exactly like this.

So to start it out, I'll try to give a basic explanation of who a nuclear reactor in layman's terms. Here goes:

The nuclear fuel is usually in the form of uranium oxide pellets. The pellets are placed in metal tubes (usually zirconium, sometimes aluminum, and much less frequently, stainless steel) that are sealed at both ends.  These are called fuel rods.  The metal wall of the rod is usually referred to as the "fuel cladding," and is the first of many engineered barriers between fuel and public.  In most reactor designs, the fuel rods are assembled into 'fuel bundles,' which are arrays of tens to possibly hundreds of individual rods.  As well as fuel rods, the bundles may contain spacer rods that only have structural material in them (no fuel), or possibly rods that are loaded with a 'nuclear poison' instead of fuel.  In this case, 'poison' does not necessarily mean the material loaded in the rod is toxic to human health, but that it has the ability to 'poison' the fission chain reaction by absorbing neutrons without an accompanying nuclear fission.  Typically used nuclear poisons are halfnium, cadmium, and boron.  The spacer rods and poison rods are used to shape the neutron flux profile across and to equalize the amount of heat production across the cross section of the reactor.  
As well as rods, the bundles have several structural elements, such as bundle end plates having channels to direct the flow of cooling water into and out of the bundle, spacer elements to prevent bowing of the rods, and edge posts that the various structural elements are attached to.  In some commercial reactors, the various rods are inserted individually into the core rather than being bundled.  The completed reactor core will contain thousands of individual rods that are arranged in a roughly cylindrical shape.  Along with the fuel rods, the core also contains control rods; these are 'poison rods' that can move into or out of the core to control the fission chain reaction, including entirely stopping the fission process, or "turning off" the reactor.  The core is placed into a reactor vessel, which is essentially a stainless steel tank with walls varying from six to ten inches of solid steel.  Along with the core, the reactor vessel also contains a lot of structure that routes the cooling water through the reactor.  The top of the reactor vessel consists of yet more structural components, motors that move the control rods, instrumentation clusters, and other good stuff.  Collectively, this part of the structure is referred to as the 'containment head,' 'closure head,' 'top hat,' or some other term of endearment specific to the reactor facility.
What points should one glean out of all this discription?  First, a nuclear reactor is an extremely complex device, and is carefully designed, configured, and constructed to enable it to attain a controlled nuclear chain reaction. 

Second, the nuclear fuel is only a very small fraction of the mass of material comprising the reactor. Third, the reactor is housed in a very substantial container that is the primary means of protecting the public from the nuclear fuel and its waste products; however, as substantial as it is, it has its weak points and vulnerabilities, especially in the closure head.

Now on to the meltdown.  Most of this information is directly from my Grandfather. Prior to the earthquake, at least some of the affected reactors were operating at power, fat, dumb, and happy.  The earthquake hit, and automatic protective systems shut down the reactors by rapidly inserting the control rods into the reactor cores.  Heat generation in the core dropped drastically, the turbines tripped off the line, and backup power supplies kicked in.  However, since 7 percent of the heat produced in the reactor comes from the decay of radioactive fission products and only time is going to decrease that production, core cooling is still needed.  The Coolant pumps are still operating, and cooling heat exchangers are easily handling that heat, and there is no evidence anywhere that any of the reactor containments have been damaged, all is still well with the world....a big earthquake occurred, and all the safeguards worked as designed.  

Now, several minutes later, along comes a 30 foot tall flood of seawater sweeping away everything it contacts...the transformer yard connecting the complex to the electrical power grid is submerged and shorted out by seawater.  The transmission lines are swept away.  The emergency diesel generators are shorted out and drenched in seawater.  The lights go out, and some feeble battery powered emergency lights come on.  The coolant pumps, which moved the decay heat from the reactor core to the emergency cooling heat exchangers stop running.  All plant instruments not provided with battery backups no longer work.  The plant operators start to realize that their world just turned brown and they are in a world of hurt.  Minutes now feel like hours as the operators try to determine the condition of their reactor plant. 

The first and foremost thought in the plant operators is that core temperatures are going to start going up toward a core damage event unless some cooling water flow can be started and maintained.  Here I'm just guessing, based on a general knowledge of various reactor plant design, but there is likely a stopgap reservoir of emergency cooling water available that can be moved into the core by air pressure from banks of high pressure air tanks.  This will provide cooling for a short interval of time in the absence of electrical power, but once that is gone, there is go good means of moving cool water through the reactor core. 

Again, I'm guessing at exact numbers, but at the start of the festivities, the water temperature in the core is probably around 600 degrees F, and the temperature at the inside of the reactor fuel is about 700 degrees F. The temperature difference between the center of the fuel and the surface of the cladding needed to remove the decay heat is less than 100 F degrees.  As long as the amount of heat being generated by the fission products can be moved from the surface of the fuel rod to somewhere/something else, there is no danger of damaging the rods or releasing fuel, since the melting point of the fuel cladding is a bit more than 3300 degrees F.  All is well (relatively) until the plant operators run out of cool water that can be moved into the core. Once that happens, the decay heat starts increasing the temperature of the water in the reactor vessel.  Even though at a temperature far above the 212 degree F boiling point of water that immediately comes to the minds of most of us, the pressure in the core is keeping it in a water state instead of steam.  Somewhere in the piping attached to the core is a steam bubble with a pressure corresponding to a 600 degree F boiling point.  

That pressure is probably somewhere around 1800 pounds per square inch (psi).  To this point, heat removal from the fuel rod has been by conduction into the water, which is a very efficient heat transfer mechanism.  The surface temperature of the cladding is only slightly (10 or 15 F degrees) greater than the water temperature.  Since heat is still being produced in the fuel by nuclear decay, the water temperature in the core now starts increasing (the heat transferred into the water has no removal route, so starts increasing the water temperature.  The hotter the water gets, the hotter the fuel temperature becomes in order to keep moving heat out to the coolant water. 

The water temperature next to the fuel rod is now higher than the boiling temperature corresponding to the water pressure, so boiling starts to occur.  This is still a very efficient mechanism for heat transfer, and the Japanese reactor plants are designed for boiling to occur in the reactor vessel (they are a class of reactors referred to as boiling water reactors).  However, since coolant is no longer being pumped through the vessel, a steam bubble now starts forming in the top of the vessel.  At this stage of the accident, core temperatures slowly increase, pressure slowly increases, and the size of the steam bubble slowly increases.  

So far they are within the systems design limits, so no serious damage is occurring...however, as the cladding temperature elevates, it starts undergoing a metal-water reaction.  In layman's terms, it starts rusting...the zirconium metal on the surface of the rod gets converted to zirconium oxide by robbing oxygen atoms from the water.  A process most of you chemistry people will recognize as oxidation. A little rust never really hurt anyone, right?  Wrong...when the oxygen is removed from the water molecule, the two hydrogen atoms left behind combine into a Hydrogen gas molecule, and start collecting in the top of the reactor vessel.  That is bad....but not as bad as events to come.

As the accident progresses, the steam bubble at the top of the core becomes large enough it starts uncovering the fuel rods..this is very bad.  The portion of the fuel rod no longer in contact with water now has convection as a heat removal mechanism, which is not very efficient.  The temperatures of the exposed region of the rods now start really ramping up.  Some of the heat produced in the exposed region is removed by conduction down the side of the fuel rod to the part of the rod still in contact with the remaining water, and some of the heat in the exposed region starts to be removed by black-body radiation.  However, as the steam bubble continues to grow, heat cannot be removed from the exposed portion of the rods as fast as it is being created, and temperature in that region continues ramping up.  The larger the steam bubble grows, the faster the temperatures increase in the exposed portions of the core.  As the exposed cladding increases in temperature, the amount of heat removed by black-body radiation increases. However, the really bad part of this is that the surface temperature needed to remove the heat being produced only by radiation is greater than the melting point of the zirconium.  As well, the hotter the clad becomes, the greater the amount of oxidation through metal-water reaction and the greater the amount of hydrogen collecting in the vapor region of the core.  Along with all this, many of the fission products in the fuel are gases.  As the fuel temperature goes up, these gasses start expanding and increasing the internal pressure of the rods.  The fuel pellets start cracking and fracturing.  The thermal stresses in the core are causing the fuel rods and structural pieces to start deforming, cracking, and otherwise turning into scrap metal.  

At some point in this nightmare, the temperature of the upper regions of the core are large enough to melt the cladding, and possibly even the oxide fuel pellets.  Meltdown has started.  The metal temperatures are great enough the cladding and other metals are not just rusting, but are actually burning.  The steam bubble in the top of the vessel is now collecting not just hydrogen, but radioactive gases from the fuel.  Molten cladding, structural material, and possibly even fuel starts flowing down into the bottom of the reactor vessel.  The bottom of the vessel still contains some water...the melt falls into that water and, since it is now thousands of degrees F greater than the water's boiling point, instantaneously boils away any water it contacts, boiling so rapid it actually acts like an explosion, and in fact this termed a "steam explosion.  It creates a pressure front that moves up the reactor vessel, continuing the process of converting the structure to scrap and knocking large pieces loose that fall into the bottom of the vessel.  These steam explosions act like a crankcase explosion: if it doesn't knock a hole through the block, it is damned well going to blow out the seals, which is what is happening in the reactor closure head.  The pressure spikes break down the seals between the vessel head parts and between the closure head and the reactor vessel.  The radioactive gasses and hydrogen are released from the primary containment.

At this point, someone is going to shout out "China syndrome."  For those of you unfamiliar with the "China Syndrome" it was term coined during the 1960's in fear of Nuclear meltdowns. The idea is that the core is now going to melt through the bottom of the vessel, then melt through the concrete floor, and then continue melting through the soil towards China (which for these reactors would be difficult, as that would require the melt to start going perpendicular to the force of gravity.  Please pardon my levity (sarcasm, actually) and bluntness of expression....but my reply is as my grandpa put it: a  resounding "bullshit."  The china syndrome makes a thrilling movie and a great boogieman for the anti-nuke crowd, but it cannot possibly happen.  Consider the core melts, the heat producing fuel is being diluted by melted cladding and structural materials, so the specific heat production is decreasing.  At the same time, the volume of the material is increasing, so there is more area for heat removal, and, far more significant, when the blob hits the bottom of the vessel, the heat transfer mechanism reverts back to conduction, which starts removing heat from the glob far faster than it is being created by the fuel.  As well, the melt no longer looks like a sphere, but becomes a plate....the ratio of surface to volume is increased a hundredfold.  The temperature of the glob drops precipitously, and very quickly is less than the melting point of the solidifies, and heat removal is great enough (the vessel makes one hell of a big heat sink) to keep it solid.  

Three Mile Island Reactor meltdown - very similar to the meltdown @ Fukushima Daiichi

At TMI (Three Mile Island), which had a very large portion of its core melt, the slag in the vessel bottom did melt about 3/8-inch into a small area of the vessel bottom....leaving it at nearly six inches of unaffected steel to contain it.  Nothing gets out of the bottom of the reactor vessel, and hydrogen production tails off to zilch.  Hydrogen and radioactive gasses have been released from the vessel, but are still within the containment barrier that exists around the reactor system.  If that containment has been breached, those gasses escape out into the secondary containment.  In reality, hydrogen is so mobile that it is nearly impossible to contain...the best the primary containment is going to do is slow down the rate it gets released.  The hydrogen gas inevitably builds up in both the primary and the secondary containment regions.  Hydrogen burns and if it builds up to more than 10% of the air volume, it explodes.  At Three Mile Island, the hydrogen buildup in the reactor building ignited before reaching an explosive mixture.  There was a pressure pulse in the TMI building, but it was not enough to breach the building walls.  At the Japanese plants, the hydrogen reached an explosive mixture, with the result of the blown out building walls and dust plume we have been seeing over and over again.  

That is the anatomy of a core meltdown.

For several reasons, there is going to be a greater release of radioactive material from the Japanese plants than from Three Mile Island.  It will primarily be in the form of radioactive gasses with some of the lighter metals (during a fuel failure, Cesium acts more like a gas than a metal)  How much material gets released to the public is going to depend largely on how well they can maintain intactness of the primary containment barrier.  In no event will the reactors release radioactives in amounts that even begin to approach the severity of the Chrenobyl accident.  That said, the fuel in the spent fuel cooling ponds is a totally different animal, and what people should be mainly concerned over.  If they melt, it goes directly into the atmosphere.  The mitigating part is that the rods in those pools have had far more time for the fission product decay, therefore likely producing a lot less decay heat relative to the rods in the reactor vessels.   

As a final word, it should be noted that the reactors at Fukushima are of the General Electric MK-I breed. As I mentioned earlier, these are boiling water reactors. Boiling water reactor design was actually pioneered in Idaho and my grandpa had a lot to do with these as well.  There are no fewer then 27 G.E. MK-I reactors in the United States that share an exact same design. Should this be a something worth worrying about, in my opinion and also that of my Grandpa's probably not. As I am sure you can see from the above literature, the circumstances that had to occur at the Japanese reactor were very particular and it would take one hell of a sequence of events to recreate the disaster. So as I always say, don't listen to the normal media B.S. about the U.S. reactor scare.