(But were afraid you might need to ask.)
Understanding events in Japan requires a fairly in depth understanding of the structures that hold the nuclear reactor and associated equipment. Looking at the pictures of the damaged reactor buildings is certainly shocking. And without an understanding of the details of how those buildings are designed and constructed it is easy to believe that the buildings and everything in them have been essentially destroyed, a belief I do not share. In this diary I will present the various layers of a nuclear containment system using the specific design of the Mark I as a model, and then review the status of each of the buildings in light of this information.
By way of qualifications, my training is in physics in which I have a bachelor's degree. I also worked for five years at Oyster Creek Nuclear Generating Station in New Jersey. Like the Fukushima plants, Oyster Creek is a GE BWR-I with a Mark I containment. It has a power rating roughly 1/3 higher than Fukushima Daiichi 1. I worked on a number of projects while there including managing the computer system used by the core engineering group to manage the reactor core and fuel usage. I also took courses in both generic nuclear power operations and also the specific operations and systems at Oyster Creek. And relative to this specific diary, I developed software used by the plant to perform required federal testing of the integrity of the primary containment structure. In the process I learned a great deal about the Mark I containment, the basis for its design and the pressures it was designed to hold. I was also a member of the plant's emergency response team and would have been one of those responding to an incident such as this.
I am also one of those 99th percentile bastards who screwed up the grading curve from time to time ... sorry about that. I understand the distinction between theory and fact, between what is known and what is conclusion. I will make every effort to keep these distinctions clear in what I present. I am also very aware that I am biased. It has nothing to do with my employment as I have not worked in the industry in over 25 years. I simply believe that nuclear is needed if we are to stop burning fossil fuels. I don't believe that 40 year old BWRs running on 50 year old technology are the answer. I believe they should be phased out as soon as practical, and watched like a hawk in the meantime. But there are far safer technologies available which I believe should be used to wean ourselves from carbon belching coal and oil plants.
In addition I have contacted an old colleague from Oyster Creek with far more impressive operations credentials than myself. He was one of Oyster Creek's core engineers and was also a Site Technical Adviser. These are engineering personnel mandated by the NRC to be present 24/7 to be immediately available to the control room operators in the event of some type of unusual event. In this capacity his job was to provide the operators with engineering advice regarding the observed symptoms, possible causes, and possible corrective actions. Had he been at Fukushima he would have been in the control room within the first minute of the earthquake. I am passing along to him technical questions which are beyond my scope of knowledge.
The whole purpose of the containment system at a nuclear power plant is to prevent nuclear material from reaching the environment. It starts with the fabrication of the fuel into ceramic pellets and ends at the walls of the secondary containment building.
The levels of containment
Fuel Pellets - The uranium fuel in prepared by mixing it with binding materials and firing it at high temperatures to create uniform cylindrical pellets in which the fuel is bound into a temperature resistant ceramic structure.
Fuel Rods - These are small diameter cylindrical tubes roughly 12 feet long in which fuel pellets are stacked end-on-end and then sealed. The material is a zirconium alloy which resists very high temperatures, much higher than those experienced during normal operations. Fuel rods are grouped into fuel bundles, however the bundle configuration does not add any significant additional shielding. (It was very high temperature interactions of this zirconium with steam that produced the hydrogen that later exploded.)
Reactor Vessel - This is the pressure vessel in which the fuel is loaded. It has stainless steel walls several inches thick and a domed cap. This is the maroon/brown cylinder in the diagram above. The reactor vessel, along with the piping that leads into and out of it, comprise the third level of containment. That piping includes the main steam line that carries reactor steam to the turbines, and the feedwater return which pumps the recondensed steam back into the reactor to continue the power production steam cycle. There are also five large pumps that recirculate water through the reactor, and other inlets and outlets for emergency cooling operations and venting.
Primary Containment - This is an enormous steel reinforced concrete pressure vessel inside which the reactor sits,along with a doughnut shaped ring at the bottom which contains a suppression pool. The suppression pool, called the torus, is easy to pick out in the diagram above. The other section is called the drywell and is the thick-walled structure that surrounds the reactor in the diagram above. It is shaped somewhat like an old style milk bottle or upside down light bulb - (Gee I wonder where GE came up with that idea.) The walls of this structure are between 4 and 8 feet thick depending on the plant's power and siting requirements. I've read that the units at Fukushima have walls that are 6 feet thick.
A variety of valves, seals, and airlocks are also part of the primary containment system. If there is a buildup of pressure inside the drywell, that pressure will also be experienced by these other components. The torus has surfaces that contact the inside of the primary containment and also the secondary containment. Given that part of its surface area contacts the secondary containment the torus is considered part of the primary containment system as well.
Secondary Containment - This is the building that surrounds the primary containment. In the diagram it is the outside building up to but not including the top floor. If you look closely you will note that the thick steel reinforced concrete walls only extend up to the floor of the upper level. Above that the walls are steel beam superstructure with sheet metal cladding. The walls of the secondary containment are not as thick as the primary containment, typically 2-3 feet. It is also not a pressure sealed structure.
I've entered the reactor building at Oyster Creek a few times in support of the primary containment integrity testing. It is just a door that you enter, no air lock, and no special clothing or protection is required to enter under normal operating conditions. There is always the possibility of radioactive contamination in the reactor building, but not to a degree that health is compromised. My total exposure from several trips into the reactor building was a cumulative 30 millirems, just over that from a chest x-ray. There is also a slight negative pressure maintained in the reactor building during operations to ensure that any airborne contaminates remain inside. Similarly there is a constant positive pressure maintained in the control room to keep any contaminants out in the even of an emergency at the plant.
Design Basis Accident
This term refers to the projected worst-case scenario - a large break loss of coolant accident, (LOCA) at full power. This is what the primary containment was designed to contain. In an accident such as this there is an immediate expansion of reactor steam throughout the containment. The containment is supposed to be large enough to allow significant expansion of the steam, and strong enough to withstand the anticipated pressure.
Federally Mandated Testing
All nuclear plant operators are required to perform periodic testing on their primary containment system to ensure its integrity. This is referred to as leak rate testing - determining the rate at which elevated pressure in the containment is lost. There are two forms of leak rate testing, local and integrated. Local leak rate tests are performed on all individual components throughout the plant which are part of the primary containment boundary. The integrated test is done by sealing the whole thing up, pumping it full of air, and seeing how well it holds its pressure. This is done by using temperature, pressure, and relative humidity measurements throughout the drywell to calculate the contained mass of air inside.
It is also necessary to adjust the volume in the calculations based on the variable level of water in the sump well. As the water level in the sump well rises, the available volume of the containment decreases. This sump level reading was at the core of criminal charges against members of the Three Mile Island management back at the time of the accident. In reviewing just about everything that had ever been done there in the aftermath of the accident, it was determined that they had altered their sump level readings in an official log for the purpose of fudging their leak rate test results.
Known Vulnerabilities of the Mark I
There are two major problems which have been identified with the Mark I containment. One of these caused a number of engineers to resign from GE in protest back in 1976. Their concern was that the containments were not being built strong enough to truly withstand the pressures in a design basis LOCA. They believe that under these extreme conditions the drywell structure itself will break open and directly expose nuclear materials to the environment.
There was also a study done by Sandia National Labs in which they determined that if a full scale core meltdown occurred in which molten core material escaped the reactor vessel it would have an 80% chance of breaching the containment. Specifically their concern was that the material would spread along the floor of the containment until it reached the walls where it would then take advantage of the wall-floor junction to eat its way out of the containment.
I do not have enough of a background in either materials science or nuclear engineering to evaluate these claims one way or the other. For the sake of prudence I presume they are correct. Even so, I am not particularly worried about these modes of failure at Fukushima under current conditions. Here is why.
The reactors at units 1, 2 and 3 all have maintained some measure of cooling water with more relief expected soon. They do not appear to be in danger of a full meltdown at this time. As such, the danger mentioned by the Sandia study should not become an issue. Provided the cooling water continues to be pumped through the reactors there should be no danger of fuel escaping the reactor vessel.
Also, the amount of power being produced by the reactors has been dropping steadily and is by now only few percent of full power. In the absence of a full scale meltdown and fuel once again producing a nuclear reaction there is just not the amount of energy available inside the reactor to create a steam explosion sufficient to pop the containment structure. As such, the scenario that caused the engineers to quit GE is no longer on the table now that power levels have significantly dropped.
There are two different venting processes that have been used at Fukushima units 1-3. In each case the venting was done to relieve pressure on the vessel in which it was contained. The first step was venting of the reactor vessel to the primary containment. This was done a number of times to reduce pressure in the reactor vessel and also as means of removing heat. The vented steam was directly pumped into the torus suppression pool where it was ejected through several outlets into the torus water. This process accomplished a few things. It cooled the steam and caused at least a percentage of it to condense back into water. It also provided some cleaning since any heavy particulates would be likely to stay in solution than bubble up out of the water and into the primary containment atmosphere. This is why I find it doubtful that significant amounts of uranium or plutonium have left the primary containment boundary. Any that left the reactor most likely sits in the bottom of the torus.
The second venting process is from the primary containment to the space just above the secondary containment boundary. (I had surmised they were venting into this space earlier in the week for several reasons. This was confirmed by one of the panel members at the conference on Fukushima held at MIT on Tuesday.) This was initially done to relieve pressure within the primary containment. It was later also done to remove hydrogen gas from the primary containment after the zircaloy fuel rod cladding began to oxidize. If you refer back to the diagram, notice that the top level is a large open space with relatively little equipment. The name of this space is technically the refuel floor, however for clarity of understanding I will be referring to it as the refuel level, and the actual flooring there as the refuel floor. You'll also note that the construction is much lighter up there. Instead of being steel reinforced concrete the walls are a steel beam superstructure with sheet metal cladding. It was into this space and not the secondary containment portion of the reactor building that the hydrogen gas was vented.
The Hydrogen Explosions
In the explosions at unit 1 and unit 3, the hydrogen venting resulted in major explosions. At unit 1 the force was enough to puff the roof up prior to it collapsing back down onto the refuel floor. At unit 3 the blast was much stronger as it was enough to blow the roof clear and mangle the steel beams. Even so, it is highly unlikely that significant damage was done to the secondary containment below. I have likened this to having a tin shack on top of a steel reinforced concrete bunker. The shack was blown away but the bunker remains intact below the rubble, shielded from the blast by the massively thick refuel floor.
You can see this in the video of the explosions as well. An explosion is actually a rapidly expanding and highly pressurized volume of gas. The leading edge of the explosion is called its shock wave. As that shock wave meets various materials that form the boundaries of its containment it has generally three things it can do. It can push on through that boundary, reflect fully, or some combination of the two. What determines the path taken is the back pressure behind the shock wave. If there is sufficient pressure pushing it then it will drive through the boundary material. If there is relatively little or no pressure behind it then it will fully reflect providing the material is sufficiently strong.
In the videos of the explosions you can see an initial pulse that pops the roof. At unit 1 the pulse was almost ghostlike and rose to about two and a half times the height of the reactor building. At unit 3 the pulse blew out the back righthand corner as it rose to about one and a half times the height of the building and was yellow-orange in color. This removed any possibility of back pressure on the other side of the shock wave impinging on the refuel floor. Instead of trying to force its way through several feet of steel reinforced concrete the shock wave reflected back into the low pressure created by the removal of the roof. This is the second wave of the explosions you see in the videos, that push of smoke from inside that comes a brief moment after the initial flash. This is where the great bulk of the energy of these explosions was directed. The secondary containments below were shielded from these explosions which happened outside of their enclosed spaces.
Unit 1 explosion
Unit 3 explosion
It is for this reason that I am highly skeptical of claims that everything in the reactor buildings has been damaged or destroyed. It is also for this reason that I am not alarmed by the photos of rubble at units 1 and 3. If you look closely at the photos and compare them to the diagram you will be able to distinguish the vertical steel beams of the refuel level superstructure. You can see these same beams in the photos of units 1, 3 and 4. At unit 4 the explosion was much less powerful and much of the wall cladding remains in place. You can see that this refuel level extends down four panels. That same level can be noted in photos of units 1 and 3 as the bottom of the pile of rubble. The reactor, secondary containment, and primary containment sit beneath the pile of rubble.
Units 3 and 4
The Fuel Pools
These are located on the refuel level and are recessed into the floor roughly -30- 45 feet [ed note: conflicting sources of info resolved]. The explosions would not have necessarily damaged them as once again, the energy of the explosion had a much easier path than through the steel reinforced walls. There do appear to be leaks in the pools at both unit 3 and 4. This is based on the rate of water drop in each pool being faster than would be the case simply from evaporation and/or boiling. Given the relative energies involved in the earthquake versus the hydrogen explosion I find it likely that the earthquake was the cause of the leak in unit 3, especially since it was almost certainly the cause at unit 4.
The hydrogen explosion at unit 4 was due to hydrogen produced in the fuel pool itself. This was only possible due to the water in the fuel pool dropping below the level of the top of the fuel rods. Also, as can be concluded from the photos of unit 4, the explosion was much less energetic than that at units 1 or 3.
Another factor that must be considered relative to the fuel pools is that they need boron in the water to prevent the fuel from starting a nuclear reaction. Dry rods next to each other cannot create a nuclear reaction because the neutrons are too fast to interact. They need water to slow them down enough to create new fissions. Boron absorbs neutrons. At the same time, the pools need water to remove decay heat from the rods. So, water + boron = very important, water alone = very bad.
Current Status of Containment Systems at units 1 - 4
[ed note: This information is superseded by information in the diary: Fukushima Status Update 3/27 (Reactor water is in the ocean) which reflects new information received as of early AM March 27nd.]
The fuel pellets and fuel rods in the reactor have definitely experienced some melting due to high temperatures and exposed fuel. Those first and second level so nuclear containment have clearly been compromised.
The reactor vessel and associated piping appear to be fully intact. Were this not so the reactor would not be able to retain the water being added and there would be unexplained pressure drops.
The primary containment also appears to be intact. It was well shielded from the hydrogen blast and has not had any other reported troubles.
The secondary containment also appears to be intact. There also have been no reported problems with the fuel pool at unit 1 despite the explosion.
The fuel containment has been compromised as is made clear by the hydrogen explosion in the torus. It is also probable that there has been fuel melting.
The reactor vessel and associated piping seem to be intact here as well. There was a brief time when they had trouble maintaining water level increases but that was apparently resolved as a valving issue.
The primary containment has damage but reports from TEPCO state that the damage did not breach the primary containment boundary and instead is in an internal section of the torus chamber.
The secondary containment has been compromised by flying debris from the explosion at unit 3. It was reported that there was a hole in the wall of the reactor building. Whether any equipment was affected by this is not clear from reports.
There have been no reports of fuel pool issues at unit 2
Clearly the fuel integrity has been compromised at unit 3 and there has been fuel melting there as well.
The reactor vessel and associated systems appear to be intact and reactor water level is able to be maintained.
There are no reports of damage to the primary containment at unit 3 and no reason to expect any.
The secondary containment at unit 3 took a hell of a hit from the hydrogen explosion, much more energetic than the others. From the various photos I've seen it still appears that the damage was confined to the refuel level. I do not discount the possibility of damage within the reactor building as a result of this blast but have not seen any evidence that would confirm this. Also, TEPCO was fully up front in announcing the damage to the unit 2 reactor building. I cannot see why they would not do the same if unit 3's reactor building were damaged.
The fuel pool at unit 3 is also a major concern. As noted earlier, it appears that there is a leak in the pool causing it to lose water faster than the heat load should cause. Many have raised concerns about plutonium due to the use of MOX fuel at unit 3. However, that only began recently and there should not yet be spent fuel in the pool containing MOX. There is plutonium in the pool as a result of it being a fission product. But there is not as much as would be the case with MOX fuel.
The condition of the reactor and containments at unit 4 is relatively unimportant given that all of the fuel was offloaded to the fuel pool to do maintenance on the reactor shroud, (the wall immediately surrounding the reactor cylinder).
The fuel pool at unit 4 may be the biggest concern at the moment. There have been reports that it had run completely dry but that appears now to have been unsubstantiated rumor. But regardless, it is losing water faster than it should and refill operations are crucial. One of the saving graces here is the fact that they don't need to get the water directly into the pools. If you refer again to the ubiquitous BWR diagram you will note that the pools are recessed into the refuel floor. When the helicopters drop water on the building, as long as it hits the floor it has a reasonable opportunity to drain into the fuel pool. Obviously some doesn't make it, but at least it is not as hopeless as it might seem to someone who believed that a direct hit was needed to get water in the pools.
Units 5 and 6
These reactors were both in cold shutdown at the time of the earthquake and did not have near the cooling needs as units 1-3. They have remained stable and there have been no reports of fuel breakdown or melting. Their fuel pools have also been out of the news, presumably a good sign.
Having off-site power back is enormous. They are now in the process of getting all of the plants there connected. This will once again make pumps and other systems available to help move water and resolve heat issues. Also, being a week out now from the initial event, the nuclear decay heat being generated is substantially lower than it was. This is because much of the decay heat arises from very short-lived fission products which have pretty much fissioned away by now. This makes the cooling requirements for the reactor cores at units 1-3 much lower than they were a week ago.
Bottom line, they're certainly not out of the woods yet but a full meltdown at any of the units seems highly unlikely at this point. Getting the situations in the fuel pools at units 3 and 4 is probably the highest priority at the moment.
[edited 5:35 to fix typo] [again at 7:02 for "zircaloy"]Updated by kbman at Fri Mar 18, 2011 at 11:34 PM PDT
I am trying to get to comments as quickly as I can. As there are quite a few coming in it may take a bit of time to get back to you. I will continue tomorrow as well so check back if you haven't seen an awaited response.
Updated by kbman at Sat Mar 19, 2011 at 03:57 PM PDT
As per an update b00g13p0p, damage to the torus at unit 2 has been confirmed. http://www.dailykos.com/...
This makes clear why they were hooking up unit 2 to the off-site power as their top priority. Continuing to remove heat from the reactor by venting steam was still necessary as long as pumps are not operating. But now that the torus is ruptured and with their already being a hole in the secondary containment, they were no longer able to retain the steam releases. Getting powered cooling restored to unit 2 was crucial to being able to stop belching radioactive steam.