Dr. Edwin S. Lyman
Senior Staff Scientist
Union of Concerned Scientists
March 3, 2006
1. MOX Fuel Relocation During LOCAs
The planned use of MOX (pluthermal) fuel in the Genkai-3 reactor in Saga Prefecture will be well outside the existing commercial experience base for MOX fuel with regard to plutonium concentration and fuel burnup. Saga Prefecture acknowledges this, but asserts that even in the absence of commercial data for the regime to which the fuel will be exposed, sufficient experimental data exists and analytical tools are reliable enough to make accurate safety assessments.
This assertion is not justified. While there is a very small quantity of experimental data on the performance of MOX fuel during normal operating conditions at high burnup, there is a huge gap in data on the behavior of MOX fuel under accident conditions. For some accidents that could result in severe consequences, experimental data does not exist to properly validate the codes used to assess the performance of MOX fuel. For example, the phenomenon of fuel relocation during a loss-of-coolant accident (LOCA) is not well understood for MOX fuel. Fuel relocation is the slumping of the fuel column that may occur following the clad ballooning stage of a LOCA. This slumping causes an increase in decay heat in the ballooned region, which may cause the peak clad temperature (PCT) to exceed safe limits during a LOCA.
Saga prefecture’s February 7, 2006 document, “Safety Issues Concerning the Pluthermal Program for Genkai Nuclear Power Plant Unit 3″, does not address fuel relocation. This new finding should be taken into account.
The French nuclear safety organization IRSN has argued that the relocation phenomenon may be more severe for MOX fuel than for high-burnup uranium fuel because the amount of RIM-like (high porosity) material that is generated during MOX irradiation is believed to be greater than the amount of RIM material generated in uranium fuel at a similar rod-averaged burnup. The reason for this is that in high-burnup uranium fuel, the RIM structure only occurs at the outer periphery of uranium fuel pellets, whereas in MOX fuel it also develops in the plutonium-rich clusters that occur in the MOX fuel generated by the MIMAS process, because of the high local burnups that occur in the plutonium-rich clusters. Since the clusters themselves are distributed across the entire pellet cross-section, MOX pellets have a greater area that is exposed to high local burnups and becomes porous than uranium pellets.
In a LOCA, it has been demonstrated that the rapid heating of the fuel pellets results in stresses that cause RIM-like regions to crack and crumble. This produces a powdery material that can collapse and fill the ballooned area of the clad. To the extent that MOX fuel has more RIM-like material than uranium fuel, the relocation effect may be more severe than in uranium fuel.
However, the existing experimental database, while suggestive, is not adequate to make quantitative estimates of the relative severity of the fuel relocation effects for MOX and uranium fuel. For this reason, IRSN proposed in 2003 to conduct a new series of experiments in the Ph?bus reactor to test this effect. However, despite appeals to the U.S. NRC, IRSN was unable to secure financial support for these tests, and neither IRSN nor any other organization has been able to do these experiments. As a result, computer codes used to simulate LOCAs in a reactor with MOX fuel in the core cannot accurately model the potential relocation effect, and therefore cannot conclusively demonstrate the safety of MOX fuel during LOCAs.
The inability of current codes to adequately simulate LOCA performance for fuels outside of the existing experimental database, such as high-burnup fuel, MOX fuel and fuels with newer cladding types such as M5 has become clear during a series of experiments conducted at Argonne National Laboratory in the U.S. over the last few years. Integral LOCA tests on high-burnup uranium fuel have revealed that clad embrittlement occurs at lower clad oxidation thickness than was previously thought, which means that current LOCA regulations are not adequately protective for high-burnup fuel. One must expect that similar surprises will occur for MOX fuel, whether at conventional burnups or the high burnups that are planned for Genkai-3. Therefore, there is an urgent need to confirm that existing LOCA regulations are adequate for new fuels such as MOX.
With respect to MOX fuel performance under ordinary operating conditions, it should be noted that in the experimental MOX fuel program in the United States, the NRC required that data be obtained from the prototype irradiation of MOX lead test assemblies (LTAs) as a prerequisite to approval of full-scale MOX use. The goal of the LTA testing is to observe fuel behavior under irradiation conditions representative of those that would be encountered during the full-scale program. The LTAs used the same design, and were fabricated with the same materials and processes, as the fuel that would be used in the full-scale program. The LTAs are being irradiated to the same burnups that would be encountered during the full-scale program in one of the actual reactors where use of MOX fuel has been proposed. The LTAs will then undergo destructive analysis in hot cells. There has been no comparable activity in Japan. There was limited LTA testing in two reactors in Japan (neither of which was Genkai-3), but not with fuel at the concentration planned for Genkai-3 or to the burnup planned for Genkai-3.
Therefore, Saga Prefecture does not have a sound technical basis to conclude that existing safety margins in the Genkai-3 reactor will be adequate in the event of a LOCA if MOX fuel is used. Also, adequate data does not exist about the performance of MOX fuel during normal operating conditions in a prototypical reactor environment. Consequently, Saga Prefecture should not give its consent for the use of MOX fuel in Genkai-3 unless integral tests are conducted with high-burnup MOX fuel under simulated LOCA conditions to assess the impact of fuel relocation and other MOX-related phenomena on safety margins, and such tests show that the relocation effect will not cause such margins to be exceeded. If, however, these tests indicate that the margins will be exceeded, then obviously MOX use cannot proceed unless the emergency core cooling systems at Genkai-3 undergo appropriate upgrades.
Without doing these tests separately, the use of MOX fuel in Genkai-3 itself must be regarded as an experiment. This means that the citizens of Saga Prefecture and neighboring prefectures that could be affected by an accident at Genkai-3 are being treated like guinea pigs in this experiment by the Japanese government.
The NRC’s decision to go ahead with MOX LTA loading was partially based on the fact that only four assemblies were involved and therefore the risk would be limited.
2. The Potential for Large Radiological Releases from a MOX-fueled
Reactor
Saga Prefecture maintains that the frequency of a severe accident that could result in containment rupture and large radiological release has been calculated for Genkai-3 and found to be one in seventy million years. As a result, Saga concludes that this type of event is something that realistically cannot be considered to occur.
This is a very important point, because of the fact that in the event of such an accident, the consequences would be much more severe if MOX fuel were in the core. For example, my calculations indicate that the number of cancer fatalities resulting from such an accident would double for a core with one-quarter core MOX fuel substituting for uranium fuel. Thus it is crucial that this type of accident be taken very seriously.
However, Saga Prefecture’s conclusion that it is not necessary to consider such an accident is unjustified. First, it is well known that the absolute frequency calculated by probabilistic risk assessment has a high degree of uncertainty, and it is not appropriate to provide the results of the central value of the calculation without providing the uncertainty bands.
Obviously, the use of MOX in Genkai-3 will increase the risk to the public from a severe accident, so the magnitude of this increase should be assessed. The U.S. requires evaluation of the consequences of severe accidents, even if they are very unlikely to occur. Also, the NRC now has a policy that if a change to a reactor license could increase the risk to the public from severe accidents, this change must be evaluated, and if the increase in risk is too great, the change will not be allowed.
Finally, even if large radiological releases from Genkai-3 are improbable as the result of an accident, they can be induced by a terrorist attack. In fact, the result of actual testing of security forces at reactors in the U.S. found that even with very good security, it is possible that terrorists could attack a plant and do enough damage to cause a meltdown and a containment failure. In the case of Japan, it is likely that terrorists would be more inclined to attack a reactor that has MOX fuel in the core because of the greater consequences of such an attack. Therefore, security should be significantly increased at any reactor using MOX, such as Genkai-3. Saga Prefecture has given no indication that such an increase has taken place. As a result, the potential for a terrorist attack to cause grave consequences at Genkai-3 after MOX fuel is loaded must be taken far more seriously than Saga Prefecture appears to be doing.
3. Limitations on the Commercial Use of MOX in France
Saga Prefecture states that the reason why France has not attempted to use MOX fuel in its 1300 MWe reactors, but only in some of its 900 MWe reactors, was that only the 900 MWe reactors were needed to absorb the supply of plutonium resulting from the reprocessing of French spent fuel. However, this ignores the fact that France has accumulated a surplus of nearly 50 metric tons of plutonium that remains unused. In fact, the French plutonium surplus continues to increase, rising from 44.2 MT at the end of 2000 to 48.8 MT at the end of 2004, according to France’s declarations of plutonium holdings to the IAEA. Therefore, the plutonium demand does not match the supply in France. Moreover, France does not now reprocess its spent MOX fuel, which would further increase the plutonium supply.
The real reason why France is not planning to use MOX in its 1300 MWe reactors (or in all of its 900 MWe reactors) appears to be the inconvenience resulting from the burnup limitations imposed on MOX fuel relative to uranium fuel, as well as complications associated with the transportation and storage of spent MOX fuel. As a result, according to a recent article in a trade publication, “EDF wants to maintain more flexibility in its current reactor fleet by concentrating MOX use in certain reactors” (Ann MacLachlan, “MELOX on path to new capacity expansion, targets Japan, EDF,” NuclearFuel, February 27, 2005).
Union of Concerned Scientists
Washington Office
1707 H St NW, Suite 600
Washington, DC 20006-3962
- References:
- A. Mailliat and J.C. M?lis, IRSN, “PHEBUS STLOC Meeting” with NRC Staff (October 23, 2003). It is on the NRC ADAMS site.
- V. Guillard, C. Grandjean, S. Bourdon and P. Chatelard, “Use of CATHARE2 Reactor Calculations to Anticipate Research Needs,” SEGFSM Topical Meeting on LOCA Issues, Argonne National Laboratory, slides at 8-9 (May 25-26, 2004).
- “Blue Ridge Environmental Defense League’s Proposed Findings of Fact and Conclusions of Law regarding BREDL Contention I”, August 6, 2004. US Nuclear Regulatory Commission Before the Atomic Safety and Licensing Board. Docket Nos. 50-413-OLA 50-414-OLA.
- “Prefiled Written Testimony of Dr. Edwin S. Lyman Regarding Contention I”, July 1, 2004. US Nuclear Regulatory Commission Before the Atomic Safety and Licensing Board. Docket Nos. 50-413-OLA 50-414-OLA.
- “Rebuttal Testimony of Dr. Edwin S. Lyman Regarding BREDL Contention I”, US Nuclear Regulatory Commission Before the Atomic Safety and Licensing Board. Docket Nos. 50-413-OLA 50-414-OLA.