On prediction of “worst case” scenarios
The blog has received a great number of questions surrounding worst case scenarios. This is not surprising given that such scenarios, with varying degrees of scientific merit, have been advanced in the media. The intent of this blog is to educate, using our best available information, and so we intend to refrain from making predictions of our own. We do, however, want to review some of the terminology used in these predictions, and describe the methods used by government agencies and scientific organizations to determine what actions must be taken to inform the public.
The term meltdown describes melting of the zirconium alloy cladding, and uranium oxide (or mixed oxide, in the case of Unit 3) fuel pellets. These two structures are the first two barriers to fission release, since radioactive fission products normally exist as either solids within the fuel pellet, gases within pores in the fuel pellet, or gases that escape the pellet but remain in the cladding. When a reactor is shut down, these fission products continue to decay, generating heat. This amount of heat is produced at first at 7% of its initial rate, and then decreases as the isotopes responsible for generating it decay away. If this decay heat is not removed by cooling water, the fuel and cladding increase in temperature.
At temperatures above 1200 C, the corrosion reaction which is constantly ongoing in the zirconium cladding accelerates dramatically. The reaction’s products include zirconium oxide, hydrogen (for more on this hydrogen, see our post “Explanation of Hydrogen Explosion at Units 1 and 3), and heat. This heat continues to both fuel the corrosion reaction, and to prevent the fuel rods from being cooled. Because of the self-catalyzing nature of this reaction, safety systems are usually actuated in such a fashion as to provide a large margin of safety to the clad reaching 1200 C.
If multiple failures prevent these actions from being taken, as was the case at Three Mile Island, the fuel rods heat up until the uranium oxide reaches its melting point, 2400-2860 C (this figure depends on the makeup and operating history of the fuel). At this point, the fuel rods begin to slump within their assemblies. When the fuel becomes sufficiently liquid, slumping turns to oozing, and the “corium” (a mixture of molten cladding, fuel, and structural steel) begins a migration to the bottom of the reactor vessel. If at any point the hot fuel or cladding is exposed to cooling water, it may solidify and fracture, falling to the bottom of the reactor vessel.
A similar sequence of events takes place if cooling to spent fuel pools is not maintained, but at a reduced rate of progression.
Breakthrough: Operating Experience and Experiment
With the fuel at or above temperatures of 2400 C, there exists the possibility that the fuel could cause damage to the reactor vessel. The melting point of the steel making up the vessel is in the neighborhood of 1500 C. In addition, the vessel in question may have been weakened by its exposure to seawater. The sodium chloride within seawater accelerates the corrosion of steels, but usually on the order of weeks or months, not days. Nevertheless, some uncertainty as to the condition of the vessel does exist.
In the event that molten corium does, as has been the case in some experiments, penetrate the lower head of the reactor vessel, it will drop onto the concrete basemat of the containment and spread out as far as possible. The interaction of corium with concrete is known to produce a buildup of non-condensable gases within the containment, a process called molten-core concrete interaction (MCCI).
In the wake of the Three Mile Island accident, a number of agencies undertook programs to determine experimentally how corium would behave when placed into contact with a concrete reactor pad. These experiments have been used to measure concrete ablation, and also the rate of generation of non-condensable gases. Over the past twenty years, these studies have focused on quenching of the corium with water.
The experiments are performed by producing a melt of un-irradiated uranium dioxide (extremely low levels of alpha radioactivity, easily avoided by the experimenters), zirconium alloy, and structural steel, in the proportions that would be present in a reactor core. This melt is sent through a nozzle used to simulate a pressure vessel lower head breach, and dropped onto concrete. Measurements are taken during the hours-long experiment using thermocouples and camera equipment, and the solidified material is examined after completion.
The experiments have shown that without water quenching, corium under conditions similar to those present at Fukushima Dai-ichi will ablate the meters-thick concrete pad at a rate of just millimeters per minute. Gases would build up within the containment at a rate which would require filtered ventilation of the containment in order to prevent rupture.
If, however, water is supplied to quench the corium as it spreads onto the reactor floor, the ablation occurs at 5-7% of the pre-quench rate, and production of gases is suppressed. The rate of ablation continues to undergo fits and starts, as the corium forms a solid crust, and then this crust is broken and re-formed.
Again, this summary is intended to explain the different pathways which molten fuel could potentially take. We do not aim to predict what’s going on in each of the reactors and spent fuel pools in question.
Analysis: How it’s done, what it means
The experiments described previously are used to validate, or confirm the results of, calculations which predict what will happen to a reactor or spent fuel pool’s fuel if it should melt down. These calculations are then used to provide the source term for an advection calculation, which predicts doses at sites removed from the plant as a function of time.
These calculations involve complex interactions between a number of different factors, such as
- The method of release: Explosive, or a slow, steady stream? Carried away by air currents, smoke, or water? How high off the ground?
- Weather patterns, both local to the site and further away
- Physical geography, both local to the site and further away
Like the methods used to model the disposition of molten fuel, these methods are validated against the best available data, which include both real-life experience like post-Chernobyl data, and the results of small-scale experiments.
The calculated doses are used by the agencies which calculate them, national and local governments to make decisions about when to evacuate or apply “take-cover” orders to people at different distances removed from the situation. Again, we recommend that our readers close to the facility heed the instructions issued by their governments.
A note about predictions of future radiation doses: in recent days a map has circulated the internet, purporting to predict high doses to the Western U.S. This map bears the seal of the Australian Radiation Service, which did not produce it. The map has been refuted by the U.S. NRC, and experts state that it more closely resembles predictions for doses after deployment of a nuclear weapon than those for a situation such as that unfolding at present.