COMMENTS ON
THE "EMERGENCY RESPONSE ASSISTANCE PLAN" FOR THE MOX FUEL SHIPMENT
FROM MOSCOW TO CHALK RIVER
Edwin S. Lyman, PhD
Nuclear Control Institute
1000 Connecticut Avenue,
NW Ste. 410
Washington, DC 20036
August 25, 2000
Overall comments
The "Emergency Response
Assistance Plan" (ERAP) prepared by Atomic Energy of Canada, Ltd (AECL)
does a wholly inadequate job of informing the public of the potential risks of
the air shipment of MOX fuel to Canada.
Numerous assertions are made regarding the robustness of both the
shipping package and the MOX fuel pellets themselves under extreme conditions
typical of an air crash --- however, no technical documents are referenced and
no other evidence is provided to support them.
On the other hand, evidence is publicly available that contradicts many
of these assertions. These statements call into question the ability of AECL to
accurately assess the potential risks of an air crash involving the Parallex
MOX fuel, and hence raise doubts about whether the ERAP that AECL has developed
is based on a sufficiently conservative "worst case scenario."
The flaws in the ERAP are extremely damaging to the
AECL's credibility with regard to its ability to safely oversee this
shipment. Without having a realistic
appraisal of the "worst-case scenario," AECL cannot make a convincing
case that it is taking the necessary precautions to ensure that the public will
be protected if the unthinkable indeed comes to pass. It should be noted that the public health consequences of the
September 1999 criticality accident in Tokaimura were greatly worsened by the
failure of government authorities and plant management to take into account a
worst-case accident in their emergency planning guidelines.
Vulnerability of the
shipping package
The ERAP contains numerous highly
misleading descriptions of the robustness of the TNB-0145/4 shipping package
that will be used for the MOX fuel air transport.
While the ERAP asserts that "the TNB-0145/4
packaging is specially designed to withstand transportation accidents,"
nowhere does it make clear that this statement refers to ground-based
accidents only. The TNB-0145/4 is a
"Type B(U)F" package, which means that it has been designed to meet
"Type B" transport standards developed by the International Atomic
Energy Agency (IAEA). These standards,
which include a series of drop tests from a height of nine meters onto an
unyielding surface (corresponding to an impact velocity of 13.2 m/s), followed
by a 30-minute engulfing fire test at
800C, have been determined by
the IAEA to be inadequate to guarantee the same level of safety when applied to
air transport as when applied to ground transport. This conclusion follows from the obvious point that the mechanical
and thermal stresses experienced by a transport package during an air crash are
likely to be considerably greater than those experienced during a typical
accident on the ground. In fact, the
IAEA assumes that Type B packages would leak when exposed to impact conditions
typical of a plane crash.[1]
As a result, the IAEA developed a new, more
stringent set of standards for packages intended for the transport of large
quantities of radioactive materials by air, known as "Type C"
standards. The Type C test regimen
involves an impact test at a velocity of 90 m/s on an unyielding surface, and a
(non-sequential) 60-minute fire test at a temperature of 800C.
These
standards, which took ten years to be developed and were heavily influenced by
IAEA member states with a commercial interest in the air shipment of plutonium,
fall far short of what would be necessary to guarantee an appropriate level of
safety for air transport of radioactive materials. In fact, in developing the standards, the IAEA itself conceded
that around 10% of plane crashes were likely to generate conditions more severe
than those represented by the Type C tests.
Nevertheless, they represent a considerably more severe accident
environment than the Type B tests. For
example, the Type C impact test delivers a kinetic energy more than forty times
greater than the Type B test.
Current IAEA regulations allow an exemption from the
Type C requirement for quantities of weapons-grade plutonium below about 3
terabecquerels (TBq), and that the amount in the Parallex shipment (about 1.4
TBq) is less than half this value.
However, this exemption value was based on the (unsupported) supposition
that a Type B package subject to a Type C impact would release from 0.3% - 3%
of its contents. For the Parallex
shipment, this corresponds to a release of 1.6 to 16 grams of weapons-grade
plutonium. IAEA would consider this an
"acceptable" release.
However, a release of this magnitude, if occurring in a populated area
such as southern Ontario, could have a significant radiological impact. In addition, there is little experimental
evidence to support releases as small as those used by the IAEA in deriving the
exemption values. It is likely that a
Type B package would lose its containment function completely in a serious
plane crash.
The ERAP asserts that "testing has shown that
the actual reserve safety margins for packages licensed to ship radioactive
materials extend well beyond the IAEA test requirements before failure,"
and infers that this applies to the TNB-0145/4 as well. However, AECL provides no references to
credible, well-documented or reproducible evidence to support this claim,
either for Type B radioactive material transport packages in general or the
TNB-0145/4 in particular.
While large, heavy spent fuel transport casks may
contain such a margin, since the thickness of their steel walls (on the order
of 25 centimeters) is determined by gamma ray shielding considerations and not
by structural requirements, there is no basis for a similar conclusion
regarding the thin-walled TNB-0145/4 package.
On the contrary, there is considerable evidence to
support the conclusion that packages are designed with little excess margin for
cost reasons. One example is the use of
elastomeric (rubber-like) lid seals instead of more expensive and
heat-resistant metal seals. Although
these seals degrade and lose their containment function after heating to
250-350C, they are still in routine
use in Type B packages because fire tests usually show that the seals remain
below this temperature after a 30-minute fire at 800C.
However, the margin to failure is not very large. A recent study by Sandia National
Laboratories (SNL) shows that for typical Type B spent fuel packages --- which
are much more massive and would heat up much more slowly in a fire than the
TNB-0145/4 --- the seal failure temperature can be reached in as little as 35
minutes at a fire temperature of 1000C, and as little as 64
minutes for a fire temperature of 800C.[2] Thus a package that passes the Type B fire
test could well fail the Type C fire test.
There is no indication that it could survive the even more severe fire
that could result from a plane crash.
AECL tries to deny this by arguing that
"packages similar to the one selected for the ... shipment have survived
tests in the 1970s based on then-current
standards [emphasis added] for aircraft flight recorders (black box) and
impact tests onto a runway at more than 200 km/hr."
There are at least two misleading and technically
unsound inferences in this sentence.
First of all, tests on "packages similar" to the TNB-0145/4
have little bearing on the TNB-0145/4 itself.
Variations in design assumptions, construction materials, and
manufacturing quality control all play a significant role in the robustness of
a package. Even two packages
manufactured from an identical design may behave differently under test
conditions. There have been numerous
examples of individual packages constructed according to approved designs
failing Type B drop tests, the most recent example being the DOE 9975 package,
a 35-gallon drum designed for plutonium transport not unlike the
TNB-0145/4. In April of this year, a
Type B drop test caused a large gap to open in the seal area of a 9975,
resulting in an effort to redesign the package.[3]
Second of all, even if one assumes (in the absence
of actual data or references) that this "similar package" indeed
survived the "black box" standards that were current in the 1970s,
then one may ask if this provides any indication that the package could withstand the standards in place today. New black box standards were introduced in
1990 because, as the Transportation Safety Board of Canada pointed out in a
letter to Transport Canada in 1995, "as recorders certified to the [then-]
existing standards failed, the standards were raised to improve the
survivability."[4]
The revised black box standards are considerably
more stringent than the IAEA Type C standards.
First of all, a black box must be able to withstand an entire test
sequence, involving impact, penetration, static crush, high or low temperature
and fluid immersion, whereas different Type C packages can be used for the
impact and thermal tests. Second, the
impact test is equivalent to a crash at a speed of 130 m/s into an unyielding
surface, about 1.44 times the Type C speed (and 10 times the Type B
speed). Third, the thermal tests are
much more severe than the Type C test.
The high temperature test involves exposure to a temperature of 1100C (typical of jet fuel fires) for 60 minutes,
while the low temperature test involves exposure at 260C for 10 hours to simulate a smoldering burn.
Given that the black box standards were developed
based on experience from actual plane crashes and are far more stringent than
Type C standards, AECL's assertion that the TNB-0145/4 would survive a plane
crash is entirely incredible. If AECL
is so confident on this point, then it should arrange to subject this package
to the current black box test sequence and invite the public to observe. A successful test would go a long way toward
convincing the public of the safety of the MOX shipment.
Dispersibility of MOX fuel
AECL's characterization of MOX fuel
as a virtually indestructible material is not supported by publicly available
information.
Its contention that "MOX fuel will not explode,
ignite or react with air or water" is a highly misleading statement. While it is true that it will not explode
and it would be difficult to ignite, there are accident conditions which
certainly could cause it to react with air or water.
Most important of these from an accident perspective
are oxidation reactions at relatively low temperatures (from 250-430C), which can result in significant
particulate formation. It has been
demonstrated that exposing sintered MOX fuel pellets to a temperature of 430C in air for 60 minutes resulted in release
of nearly 70% of the fuel in the form of particles with diameters less than 25
microns, of which over 6% was observed to be of respirable size (below 10
microns).[5] An accident involving an impact which
breaches the fuel cladding, followed by a relatively low-temperature fire,
could cause the package seals to fail, oxygen ingress into the package, and
fuel pulverization.
AECL also claims that for MOX fuel, "high
energy impact tests do not generate a significant portion of the fuel as a fine
powder that could be dispersed in an accident." However, ceramics are brittle materials and will indeed pulverize
if subjected to the impacts typical of plane crashes. According to a correlation for uranium fuel used by DOE, an
impact at 130 m/s would cause 1.7% of the initial fuel mass to be released in
the form of respirable particles, hardly an "insignificant portion."[6] This corresponds to about 9 grams of
plutonium for the Parallex shipment. If
the impact were followed by a low-temperature fire as described in the previous
paragraph, the production of respirable particles would be considerably
greater.
Conclusion
The ERAP should be rejected in its current
form. It should be rewritten, taking
into account the most recent analyses and experimental evidence regarding the
performance of Type B shipping packages and the dispersibility of MOX fuel in
air crashes. A realistic worst-case
scenario should be explicitly defined, the unmitigated consequences should be
modeled, and the impact of proposed emergency planning measures on reducing
those consequences should be assessed.
Only then will Transport Canada have a basis for assessing whether the
ERAP fully and credibly meets its requirement that it address "accidents
in which the MOX fuel samples may be released outside of the ... package as a
mixture of ceramic pellets and dust."