ML20210P693
| ML20210P693 | |
| Person / Time | |
|---|---|
| Issue date: | 08/25/1997 |
| From: | Collins T NRC (Affiliation Not Assigned) |
| To: | Gray R AFFILIATION NOT ASSIGNED |
| References | |
| NUDOCS 9708270299 | |
| Download: ML20210P693 (19) | |
Text
t August 25, 1997 o
Mr. Robert Gary Gary Research 2211 Washington Avenue (#301)
Dear Mr. Gary:
We have received your 1rtters of March 11 and 12,1997, regarding the possible burning of mixed plutonium oxide and uranium oxide fuel (M0X) in power reactors licensed by the U.S. Nuclear Regulatory Commission We have only begun to look into the issues that might arise from this(N C ).
new interest in M0X consumption; however, your letters focused on several major questions, and we would like to address them briefly. Our responses are based on preliminary information provided by the Department of Energy on the disposition of excess plutonium and our experience with M0X in commercial reactors in the 1970s.
Your major questions seem to be:
1.
What is the quantity of plutonium in a M0X-fueled reactor in comparison with a V0, fueled reactor?
l 2.
Can plutonium metal become segregated from the fuel, which is in oxide form, and accumulate in one location?
2 f'g, Ny I.>3 3., n,,_
3.
Does " crud" in the reactor piping contain plutonium?
4.
Can a nuclear explosion result because of this plutonium?
gW gg Wewillanswereachofthesequestionsinanapproximatemanner,ab b#
I requested.
(1) Quantity of Plutonium You asked about the amount of Pu-239 in the proposed M0X reactor.
Based on information provided to us by DOE, the design will be such that about i
hgg )
1 ton of Pu-239 will be processed per year per reactor. Assuming a 1 year t
operating cycle length, and thus 1 ton inserted each year, this does not result in an average of 3 tons of Pu-239 in the reactor, as you suggest.
At
\\
any one time, there are varying amounts of burnup of the fresh and previous I
cycle Pu-239, depending on the fuel power history.
This variation might t \\
typically result in the peak over a cycle of about 1.5 tons in the reactor, t
(This assumes an equilibrium cycle.
Initial cycles would build up to this with lower initial enrichments.)
According to DOE, the Pu-239 is in oxide form, Pu0, and mixed with Urss0, and is thus similar to the current U0 fuel systems.
IhePu-239willhavean 3
enrichment range of 2 to 5 percen,t, similar to present fuel.
A change to a M0X design would have current or similar fuel assembly designs and about the same weight of Pu-239 current core designs (plus U-235 and U-238 as the U-235/U-238 content of about 100,000 kg, or 100 tons, at beginning of cycle.)
cg(C,49
,i,i 9708270299 970825 PDR ORQ NRRA PDR q-
Mr. Robert Gary 2
(NotethatthecoredesignwouldincludeasignificantamountofU-235aswell as Pu-239 to provide appropriate power distributions and reactivity control.)
It will also have Pu-239 plus U-235 weight similar to current cores, since Pu-239 and U-235 provide the same energy per fission, and require the same amount of fuel to provide the same cycle energy.
The Pu-239 content, assuming 3 percent enrichment, bout I ton.and Pu-239 loaded into 1/3 of the core (a typic d design configuration) is a Enclosed for reference, is a representative graph of production and consumptlon of Pu and U isotopes in a typical UO, reactor as a function of I
fuel burnup. This covers about three 1-year cycles.
given reactor would be more complex bec(ause o)f cycle chan(Actual graphs for a ges.)
Typical fuel batch burnup, is about 40 gigawatt-days (GWD) per metric ton of fuel (MTU),
and fuel pin burnup limits are about 60 GWD per MTV.
This tends to set-enrichment limits. A M0X reactor would have a similar gra)h, with the U-235
._ curve-replaced with a similar Pu-239 consumption curve.
T1e VO graph-indicates that about 500 kg of Pu-239 is produced in current re, actors and this continues to exist for more than half of the cycle.
(Note, for the graph, there are about 100 MTU in the reactor.)
The mass of 31utonium or uranium required to reach or exceed a critical state de) ends on tie amounts and configuration of the fissionable materials and any ot1er materials present which might act as neutron reflectors, moderators or absorbers.
Information on critical masses associated with various configurations can be found in the government report, " CRITICAL DIMENSIONS OF SYSTEMS CONTAINING U-235, Pu-239, and U-233. TID-7028" USAEC 1964.
You have asked about conditions after the Three Mile Island (THI) accident. to this letter reproduces some pages from a post-event analysis of the core conditions giving the type of information you have recuested, it should be noted that the TMI-2 plant was in its first. cycle anc operated for only:3 full-power months. -Thus plutonium content was relatively low, about 130 kg compared to about 4000 kg of U-235 (See Enclosure 1 in which the THI-2 plant would have a burnup of only.3 GWD/MTV.) The Pu probably was not very noticeable in the debris, it should also be noted that considerable reactivity control material and other structures were in meltdown regions.
This would greatly increase plutonium / uranium critical mass requirements.
Burnable poison and. control rods are associated with fuel in all current and proposed power reactor designs.
(2)
Plutonium Seareaation Plutonium metal is extremely hard to separate from uranium and fission products, even under the carefully controlled conditions of an industrial reprocessing plant or an analytical laboratory.
There is virtually no chance that plutonium could be separated during an uncontrolled accident.
- Further, we are not-aware of any physical mechanisms for separating the U0, and the UO f.
Chemically, to separate the pldonium metal from the oxygen and the Pu0 likely.there would-have to be a preferenth1 reduction of the Puo,, which is not I
s
\\
Mr. Robert Gary 3
(3) Crud The term " crud" in reactor operations is normally used to describe adherent deposits that accumulate on )iping and fuel rods. These deposits are made up of corrosion products from tie oxidation of metal surfaces.
This is not a sludge-like material as you suggest. Neither the crud nor any loose debris would likely contain significant plutonium, which can normally come only from inside the fuel rods.
as loaded, could produce (Tramp U-238, which may be on the outside of the fuel Pu-239 outside of the fuel; however, the resulting Pu-239 would normally be, at most, at very low levels compared to criticality requirements.) To get plutonium out of a fuel rod, the cladding would have to fail in a major way.
Such fuel failures are uncommon. Current fuel pin defect rates are extremely low, with only 1 or 2 leakers per reactor, and many of these are holes or tight cracks that leak only gas.
Fuel leakage is monitored and limited. Thus the quantity of fuel lost frorr fuel rods in current practice is very smaII, and is very far from even one critical mass.
(4) Nuclear Exolosion A nuclear explosion in a U.S. reactor is not a credible event. While a water-steam over-heating pressurization is conceivable, it is impossible in the environment of the accident you postulate to produce a nuclear explosion. A nuclear explosion requires the yary quick bringing together of a mass greatly in excess of critical mass and the absence of neutron absorbers.
For your scenario these conditions are not realized.
(In response to your request Mr. Carl Berlinger has read your letter and this response and has no additional comments.)
Sincerely, Timothy E. Collins, Chief Reactor Systems Branch
_ Division of Systems Safety and Analysis Office of Nuclear Reactor Regulation
Enclosures:
As stated DISTRIBUTION:
File Center-RZimmerman TMartin Public WTravers KBohrer (YT #0970031)
SCollins GHolahan PMagnanelli (YT #0970031)
FMiraglia JLyons LPhillips AThadani HRichings HRichings R/F
- see-previous concurrence page EDITED BY:
R. Sanders DAT 5/1/97 DOCUMENT NAME: G:\\LTRGARY.HR3
/ ',
SRXB:DSSA SRXB:DSSA SRXB:DSSSA D:DSSA ADTl R
D:
R' HRICHINGS*
LPHILLIPS*
JLYONS*
GHOLAHAN*
T N
4 S
5/13/97 5/13/97 5/13/97 5/17/97 8/g/97 V/97
Mr. Robert Gary 3
(3) CrXd The term " crud" in reactor operations is used to describe adherent deposits that accumulate on piping and fuel rods. These deposits are made up of corrosion products from the oxidation of metal surfaces.
This is not a sludge-like material as you suggest.
Neither the crud nor any loose debris would likely contain significant plutonium, which can normally come only from inside the fuel rods.
(Tramp U-23B, which may be on the outside of the fuel as loaded, and could produce Pu-239 outside of the fuel, would normally be, at most, at very low levels compared to criticality requirements.)
To get plutonium out of a fuel rod, the cladding would have to fail in a major way.
l Such fuel failures are very uncommon.
Current fuel defect rates are extremely low, with only 1 or 2 leakers per reactor, and many of these are holes or tight cracks that leak only gas.
Fuel leakage is monitored and limited.
Thus, the quantity of fuel lost from fuel rods in current. practice is very small, and is-very far from even one critical mass.
(4) Nuclear Exolosion1
\\
A nuclear explosion in a U.S. reactor is not a credible event.
We normally do not do calculations of explosive events in reactors and are therefore, not in a position to comment on theslevel of a nuclear explosion. While a water-steam over-heating pressurization is conceivable, it is impossible in the s
environment of an accident you postulate to produce a nuclear explosion. A nuclear explosion requires the yACy quick bringing together of a mass greatly in excess of critical mass and the absente of neutron absorbers and s
predetonation neutrons. For your scenario these conditions are not realized.
(In response to your request Mr. Carl Berlinger has read your letter and this response and has no additional comments.');
/x Sincerely, s
JamesE\\Lyons,ActingChief Reactor Systems Branch Division of Systems Safety and Analysis Office of Nuclear Reactor Regulation
Enclosures:
\\
As stated
's s
DISTRIBUTION:
A File Center
< RZimmerman TMartin
'N Public WTravers KBohrer (YT #0970031)
SCollins GHolahan PMagnanelli (YT #0970031)
FMiraglia JLyons LPhillips s
AThadani HRichings HRichings R/F s
EDITED BY:
R.\\ Sanders DATED:
5/1/97
'*see previous concurrence page i
DOCUMENT NAME: G:\\LTRGARY.HR2 J
SRXB:DSSA SRXB:DSSA SRXB:DSSSA D:DSSAT ADT:0NRR D:0NRR HRICHINGS*
LPHILLIPS*
JLYONS*
_GHOLAHAN*
TMARTIN SCOLLINS 5/13/97-5/13/97 5/13/97 5/17/97 5/ /97 5/ /97 4 b \\"\\
s
R. Gary 4
/
relevant to your question.
It is relevant to your question, howev'er, that a nuclear explosion is highly unlikely.
While a water-steam over heating pressurization is conceivable, it is very difficult in the environment of an accident you postulate to produce a nuclear explosion.
A nuclear explosion requires-the yn y quick bringing together of a mass greatly in excess of critical mass, without any neutrons around to start a premature " dud" separation of the materials, and then the introduction of a neutron source at peak excess reactivity as a trigger.
There would be many more neutrons present in the conditions of your scenario from e.g., Pu-240 (from non-fission captures in Pu-239 during operation), reactions with gamma radiation and any water in the systems, and gamma reactions with fission products that would preclude a nuclear explosion.
it is a)so unlikely that a nuclear ex)1osion could occur with a moderated C.M.,'since the necessary slow neutrons ta <e too long to slow down. There is also' bound to be other debris in contact with any Pu-239 as with the TMl-2 plant,/which would act as a control poiso.
(in response to your request Mr. Carl Berlinger has read your letter and this response and has no additional comments.)
Sincerely, James E. Lyons, Acting Chief Reactor Syshms Branch Division of Systems Safety and Analysis Office of Nuclear Reactor Regulation
Enclosures:
As stated DISTRIBUTION:
/
File Center RZimmerman TMartin Public WTravers KBohrer (YT #0970031)
SCollins GHolahan PMagnanelli (YT #0970031)
FMiraglia JLyons LPhillips AThadani HRichings HRichings R/F
)
EDITED BY:
R. Sanders DATED:
5/1/97
- see previous concurrence page J
DOCUMENT NAME: G:\\LTRGARY.HR2 M
SRXB:DSSA SRXB:DSSA SRXB:DSSSA D:DSSA D ONRR D:0NRR HRICHINGS*
LPHILLIPS*
JLYONS*
GHOLAHAN*
1 MARTIN SCOLLINS 5/13/97 5/13/97 5/13/97 5/17/97 5/ /97 5/ /97
R. Gary 4
relevant to your question, it is relevant to your quettion, however, that a nuclear-explosion is highly unlikely. While a water-steam over,-heating pressurization is conceivable, it is very difficult in the environment of an accident you postulate to produce a nuclear explosion. A nuclear explosion requires the y_tr.y quick bringing together of a mass greatly in excess of critical mass,-without any neutrons around to start a premature " dud" I
separation of the materials, and then the introduction of a neutron source at peak excess reactivity as a trigger. Note that the first plutonium bomb could not be designed the same as the uranium bomb mass into another at high speed) because there(a " gun" shooting one uranium l
were too many neutrons around from the (small) Pu-240 coi. tent of the proposed device.
A faster and much l
-more complex design was required.
There would be many more neutrons present in the conditions of your scenario than in the plutonium bomb, from e.g.,
Pu-240 (from not-tission captures in Pu-239 during operation), reactions with gamma radiation and-any water in the systems, and gamma reactions with fission products.
It is also unlikely that a bomb could be made with a moderated C.M., since the necessary slow neutrons,take too long to slow down.
There is also bound to be-other debris in contact with any Pu-239 as with the TMI-2 plant, which would act as a control poison.
(In response-to your request Mr.
Carl Berlinger has read your' letter /and this response and has no additional comments.)
Ny J
Sincerely.
l Samuel J. Collins, Director Office of Nuclear Reactor Regulation
Enclosures:
/
As statu
/
DISTRIBUTION:
File Center
/
RZimmerman TMartin ',
Public Wiravers KBohrer (YT #0970031)
SCollins GHolahan PMagnanelli'(YT #0970031)
FMiraglia JLyons LPhillips AThadani HRichings HRichings R/F c
EDITED BY:
R. Sanders DATED:
5/1/97
- see previous concurrence page DOCUMENT NAME: G:\\LTRGARY.
SRXB:DSSA/
SRXB:DSSA SRXB:DSSSA ADT:0NRR D:0NRR HRICHINGS*
LPHILLIPS*
JLYONS*
'LAHAN THARTIN SCOLLINS
-5/13/97/
5/13/97 S/13/97
/p/97 5/ /97 5/ /97 22 pl Y
4 J
4 limited.
Thus, the quantity of fuel lost from fuel rods in current practice is very small, and is very far from even 1 C.M.
/
(4) Nuclear Explosion WenormallydonotdocalculationsofexplosiveeventsinIoactorsandare therefort, not in a position to comment on the level of.a nuclear explosion relevant =to your question, it is relevant to your question, however, that a nuclear explc* ion is highly unlikely.
While a water-steam over-heating pressurization is conceivable, it is very difficult /in the environment of an accident you postulate to produce a nuclear explosion. A nuclear explosion i
l requires the yany quick bringing together of a mass greatly in excess of critical mass, without any neutrons around to start a premature " dud" separation of the materials, and then the introduction of a neutron source at i
peak excess reactivity as a trigger Note that the first plutonium bomb could not be designed the same as the uranium bomb mass into another at hlqb speed) because there(a " gun" shooting one uranium were too many neutrons around from the (small) Pu-240 content of the proposed device. A faster and much more complex design was required.
There would be many more neutrons present in the conditions of your scenario than in the plutonium bomb, from e.g.,
Pu-240 (from non-fission captures in'Pu-239 during operation), reactions with I
gamma radiation and any water in.the systems, and puma reactions with fission products.
It is also unlikely that a bomb could be made with a moderated C.M., since the necessary slow neutrons take too long to slow down. There is also bound to be other debris in contact with any Pu-239 as with the TMI-2 i
plant, which would act as a control poison.
/
Sincerely, J
/
SamuelJ.Cbilins, Director
/
Office of Nuclear Reactor Regulation
Enclosures:
/
As stated
/
s DISTRIBUTION:
/
file Center RZimmermar TMartin Public
/
WTravers KBohrer (YT #0970031)
SCollins j
GHolahan PMagnanelli (YT #0970031)
FMiraglia
/
JLyons LPhillips Alhadani HRichings HRichings R/F EDITED BY:
R. Sanders D:DSSAt [ DATED:
5/1/97 ThENTNAME: G1LTRGARY.HR2 SRXB:DSSA JSSA SR SSSA ADT:0NRR D:0NRR HRICHING
/)\\ 5//3/97 LPHI LIPS JL GHOLAHAN TMARTIN SCOLLINS 5//3 /97 5/f 3 /97 5/ /97 5/ /97 5/ /97 22
ENCLOSURE 1 o
9 0
8 4
3 H
D 2 7 y
8 E
E h6 12 0 0
h o
-5 o
16 9 p/-
\\
94
/
20 I
Pu239 /
u-o
- ~/
2 3
/
9 24 9
/
U.238 g
t,
--+
s
$2 28 O
j Pu241 8
a.
4 1
/-
32 0
/
\\N l
l l
l l
l 36 0
4 8
12
- S 20 24 28 32 36 40 BURNUP. (GWD/MUT) l 1
DJCLOSURE 2
/'
THi 2 CORE MATERIALS AND f!SSION PRODUCT INVENT 0 rya Doug.las W. Akers Richard K. McCardell Malcolm L. Russell Getachew Worku*
Idaho National Engineering Labnrators EG&G Idaho, Inc.
ldaho Falls, ID 83415
, Grove Engineering Inc.
Rockville, Maryland ABSTRACT Examinations have been performed to characterize the distribution of core materials and fission products at the damaged Three Mile Island Unit 2 (THI-2) reactor.
The purpose of this paper is to summarize the current status of the core niaterials and 'ission product inventory evaluations at TH1 2 and to present results that may affect reactor source term issues.
Principal subjects include the ralocation of core materials outside the reactor vessel, the formation of complex, core material interaction products, cesium retention in prior molten material, and the retention of tellurium, antimony, and ruthenium ir 2ssociation with core materials.
Material examinations have now been performed on samples from throughout the reactor building and vessel.
These examinations have helped define the end state core configuration and materials behavior. The examination results have been used to estimate the redistribution and inventories of core materials and fission products in the damaged reactor core for comparison with expected materials behavior, particularly for fission products.
INTR 00VCT10N The Three Mile Island Unit 2 (TMI-2) pressurized water reactor underwent a prolonged loss of coolant accident on March 28, 1979, resulting in severe damage to the reactor core.
As a consequence of the TMl-2 accident, numerous aspects of light water reactor (LWR) safety have been questioned, and the U. S. Nuclear Regulatory Commission (NRC) has embarked on a thorough review of reactor safety issues, particularly the causes and effects of severe core damage accidents.
The nuclear community acknowledges the importance of examining TMl-2 in order to understand the nature of the core damage.
Immediately after the accident, four organizations with interests in both plant recovery and acquisition of a.
Work supported by the U. S. Department of Energy, Assistant Secretary fo-Nuclear Energy, Office of LWR Safety and Technology under DOE Contract No. DE-AC07 761001570.
-:73-
- 46 W TW4&f 499pWs-4,g,g.g,,
accident data formally agreed to cooperate in these areas.
These organizations [ General Public Utilities Nuclear Corporation (GPU Nuclear owner / operator of THI), Electric Power Research Institute (EPRI),
the NRC, and the U. S. Department of Energy, collectively known as GEND) are presently involved in postaccident evaluations of THl 2.
The DOE is providing a portion of the funds for reactor recovery in those areas where accident recovery knowledge will be beneficial to the U. S. LWR industry.
In addition, DOE is providing funds to acquire and examine sample >
obtained from the damaged reactor core.
The examinations of the THI 2 reactor core began shortly after the l
accidentwiththeevaluationofplantmonitopinginstruments,and sampling and analysis of the reactor coolant.
This was followed by examinationsofsamplesofhighlycgnjaminatedwater,andsedimentfrom the auxiliary and reactor buildings *. Much of this data has been analyzedandprovidessufficientinformationtocharactgrizethe distribution of core materials outside the reactor core.
asdeterminedbyclosedcircuittelgvision(CCTV),fugtaftertheaccident The end state configuration of the THI 2 core mechanical is illustrated in Figure 1.
The probing, and core boring operations initial examinations for core materials were performed on sections of control rod drive leadscrews, which e nded from the upper plenum through 9
to the beginning of the core proper Visual observation: through the leadscrew guide tubes indicated the presence of a void region above a October 1983gefueldebris(upto94cmdeep)whichwasfirstsampledin region of lo As part of these examinatior.s. it was determined that a crust layer was present below the debris bed. This crust layer was finally sampled using specialized core boring equipment adapted from the mining industry.
The visual examinations of the lower region of the reacter core (reference 8) indicated the presence of a central prior molten region contained by partly or fully gtallic crust layers. Destructive examination of the core bores obtained from the lower core indicate that the crust layers were composed principally of metallic constituents of the core that had not oxidized during the accident.
Video examinations of the region below the reactor core were also performed, which indicated that prior molten material had relocated to the lower head of the reactor.
Samplesofthisdebriswereobtainedthroughtheannulusbetweentg thermal shield and the reactor vessel.
Examination of this debris indicated that it was a mixture of prior molten fuel, cladding, and structural materials which had oxidized and flowed down onto the lower head of the reactor.
These examinations provide the bulk of the data on the distribution of core materials in the damaged reactor core.
The following sections discuss the distribution of the bulk debris in the reactor vessel, the segregation and redistribution of the elemental constituents of the various core components, and finally, a discussion of the redistribution of fission products and the possible effects on source term issues.
-;74--
~
--e~.'+
'QL
9
'N -
S, ora +n._g-a-
^
1, x
a w
s u -a,
)_.-
S'
((
T' s
28 inlet l
1 A Inlet i
I 0
0 0
_ 0 0
l lUM
))
0 0
0 :
,:0 1) : 0: 0, 0,: 0,: 0 l
J 0
,. ; ey F
'- 0 3
(. ij*
Upper grid
(
e,pf,g Y
damager -
il ypa%
II Coating of previously.
-- Loose core debris a
-L surfaces
.(" M-
$glI$$
1l
- - Crust molten material on 1
bypass region interior ld 6
i ls 9
7 reviously molten P
g-H'
]
material f
l Hole in I
baffle plate
}
D 3
\\
(($h$hSh Ablated incore s,
r Lower plenum debris instrument guide i
0 #
2'agT W 7
~
Possible region
~
depleted in uranium
~
I l
9.
- -sors Figure 1. TMI 2 end state core condition.
i
-275-
CORE MATERIALS INVENTORY
/
The original core materials inventory included approximately 94,000 kg of U0 fug*gd35.500kgofcladding, structural,andcontrol 3
materials As a consequence of the accident, some material has been added to this inventory due to ablation of other reactor vessel materials and oxidation of some of the metallic' structural components.
Approximately 229 kg of structural steel was estimated to have been ablated from the underside of the upper plenum and 182 kg from baffle plates, the former core wall, and other structural features not in the core but in the reactor vessel.
In addition, some of the metallic magrials in the reactor core were oxidized producing approximately 459 kg Hp This would result in a total addition of oxygen as oxides to the core of about 3300 kg.
Consequently, the total mass of core materials is about 133,250 kg.
Of this total, about 100 kg was relocated outside the reactor vessel during the accident.
GPU Nuclear has removed much of this external debris as part of the defueling operation.
Table I lists the postaccident distribution of core materials at TH1-2.
The largest fraction of the core materials (33%) is located in the partial fuel rods in the periphery and at the bottom of the reactor core.
Examination of these damaged assemblies indicate that they have not lost any of their core materials inventory and were not subjected to high temperat.ures.
The remaining core material repositories range in composition from the prior molten fuel and structural materials in the central core consolidated region to the mixture of intact and previously molten materials in the upper core debris bed, The central consolidated mass region is composed of a core of prior molten fuel materials surrounded by layers of crust material with differing compositions. Analysis of the data in references 8 and 13 indicates that the upper crgst is composed of 2450 kg of debris with an averagedensityof8.3g/cm,andthelowercrugtiscomposedof8760kg of material with an average density of 7.3 g/cm.
These nominal values have associated uncertainties of 30 40% due to the heterogeneity and distribution of the debris in the crust layers.
This results in a total of 25,990 kg of prior molten debris between the crust layers.
The upper debris bed (20% of core mass) is composed of a mixture of relatively intact fuel materials and prior molten fuel structural and control material.
This part of the debris bed is de up of relatively friable material which has a density from 3 5 g/cm Intact cladding shards and control material fragments were present in the debris.
The prior molten material relocated to the lower reactor vessel head (19,100 kg) has been examined only at the surface of the debris bed.
This material (reference 19), which may not be representative of all material on the lower reactor vessel head, indicate that the debris is a homogeneous mixture of fuel materials (V,Zr) with relatively small amounts of structural and control materials.
Nondestructive examinations of the
-2n-
TABLE 1.
ESilHATED POSTACCIDENT CORE MATERIALS DISTRIBul10N Estimated Uncertainty
- Percent of Core reaion guantityfka)
(%)
Total coref%)
Intact fuel assemblies (Partially or fully intact) 44,500 5
33.4 Central core region resolidified mass 32,700 5
24.5 Upper core debris bed 26,600 5
19.9 Prior molten material on the Lower reactor vessel head-19,100 20 14.3 b
Lower ccre support assembly 5,800 40 4.3 b
Upper core support assembly 4,200 40 3.2 Outside the reactor vessel 100 c
0.3 1
a.
The uncertainty estimates are based on defueling.
Those areas which have been defueled at this time have relatively low uncertainties [,)
whereas those which have not have relatively high uncertainties, b.
The lower core support assembly is that portion of the reactor vessel below the core which includes the lower grid assembly and five flow distributor plates.
The upper cQre support assembly is a coolant flow region outside the vertical baffle plates [ ) which make up the peripheral boundary of the core.
c.
Estimates of the amount of fuel material outside the reactor vessel are based on nondestructive evaluations of reactor components in the reactor and auxiliary buildings.
They range from 60 to about 430 kg.
~217-m.,_
/.
/-
t lower head suggest that st areaofthereactorvesselgeturalmaterialsmayhaverelocatedtothis
, and examinations are in progress to better define the composition of this debris.
The remaining repositories in the reactor vessel (10,000 kg) have ret been characterized, but are expected to be similar in composition to the material found on the surface of the debris bed on the lower reactor vessel head.
CORE MATERIAL DISTRIBUTION Three principal types of core materials are present in the THI 2 reactor core: the fuel rod constituents (uranium, zirconium, and tin), the control rod materials (silver, indium, and cadmium), and the structural materials (stainless steel and Inconel).
Examinations by GPU Nuclear indicated that only about 100 kg of core material have relocated outside the reactor vessel, as previously discussed, and that the bulk of this material, a mixture of prior molten and intact fuel and structural materials, is located in the reactor coolant system in the steam generators and in the pressurizer.
All other core materials remained in the reactor vessel.
Fuel Rod Constituents Before the accident, the THi-2 fuel rods consisted of 94,000 kg of 00, 24,000 kg of zirconium present mostly in the form of metallic 2
Zircaloy, and approximately 370 kg of tin as a component of the Zircaloy.
Tne U0, present as the oxide fuel, may be dissolved by the zirconium, 2
which can be oxidized by steam as part of the high temperature melting process.
During the accident at THI-2, a significant fraction of the metallic zirconium was oxidized to Zr02 as indicated by the production of hydrogen. Approximately 43% of the Zr was oxidized during the accident; this estimate is based on the amount of H2 (459 kg) produced by oxidation. However, some of the oxygen produced would have been used to oxidize structural metallic material (e.g., Fe, Cr).
Consequently, the estimated 43% oxidation of Zr is conservatively high, Tin is expected to THI-2g9,etallicintheconditionsaresentduringtheaccidentat remai m and its behavior during tie accident would h expected to be different from the oxidized zirconium.
Analyses were performed on samples obtained from throughout the reactor core, which provided information on the redistribution of uranium in the reactor system.
Table 2 lists the distribution of the principal fuel rod constituents in the various regions of the damaged reactor core.
As indicated in the table, this distribution accounts for approximately 97% of the total uranium inventory.
The principal repositories for uranium are the partial fuel assemblies, the upper core debris bed, debris relocated to the lower reactor vessel head, and the debris in the central consolidated region.
The partial fuel assemblies around the periphery of the core (22.7%), and the partial assemblies below the central consolidated region (10.7%) make up the largest repository of core
-078-
. /
a/
g[
TABLE 2. FUEL MATERIAL DISTRIBUTION IN THE REACTOR VESSEL s
Core material distribution Percent of Inventory a Core Material Recository Uranium Zirconium Tin Upper reactor plenum
-b-
-b-b-
Upper core debris 24 13
-c-Upper crust region ceramic 1.3 1.2 2.3 metallic 0.3 6.1 Consolidated region ceramic 12 18 metallic 0.2 5.8 Lower crust region l
ceramic 3.6 2.8 9.3 L
metallic 5.6 26 s
l Intact fuel rods 33 33 33 Lower reactor vessel head 15 11
-c-Lower core support assembly 4.6 3.3
-c-Upper core support assembly 3.3 2.4
-c-Total 97 91 82 a.
Percentage of the total amount of the element originally present in the core.
b.
Insignificant amount (<0.1 wt%) based on the upper plenum measurements.
- c.. Elemental constituent not detected based on detection limits of approximately 0.1 wt%.
k l
-219-
__s
ri materials following the: accident.
Examinations indicate.that the intact portion of these-feel rods'show no evidence of accident damage.
3 The next largest repository for uranium at THI 2 is the loore debris in the upper part of the reactor core.
This material is particulate debris from 1 to 5 mm in diameter and is composed of prior molten fuel materials mixed with partial fuel pellets, cladding, and structural material pieces.
This debris averages 75 wt% uranium which is higher than the core average concentration of abcut 66 wt%. The hign uranium u
concentration is due to the melting and downward relocation of zirconium to -the central part of the reactor core which is discussed in a following l
paragraph.
Approximately 15% of the core inventory of uranium was relocated to the lower reactor vessel head.
Examinations performed on debris from the surface of the debris bed suggest that this material is a relatively homogeneous mixture of uranium and zirconium concentration (65 wt%) near the core average,, with the uranium o
The consolidated core region, compond of-the upper and lower crusts and the central core region, makes up the bulk of the remaining core 1
inventory-(17%). The uranium content _is quite variable due to the high degree of heterogeneity of the materials in this part of the core.
The average concentrations fo uranium are: upper crust (49 wt%), central region (54 wt%), and lower crust (34 wt%), with a significant range of concentrations at all locations due to the heteroganeity of the debris.
The lower crust contains only about half the expected amount of uranium as compared to the core average, indicating the presence of significantly more structural components in this part of the core.
In geaeral, these data suggest a very distinct segregation between the U-Zr-0 and the metallic or oxidized structural materials. This is consistent with the-expected behavior of uranium, in' that it is expected to interact significantly only with the 7.irconium in the system. Structural matericls are-present as contaminants [and are] located in inclusions or at grain boundaries:in the U-Zr-0 matrix.
The zirconium date in Table 2 show a similar distribution to that indicated for uranium, except that the percentages of inventoryire less, reflecting differencas between the chemical behavior of the ceramic 00 and the metallic zirconium. The dah indicate that almost half tne 2
zirconium originally present in the Jpper Core (i.e., the upper debris bed) has relocated to lower regions of the reactor core.
The data from the crust layers and the central-core region indicate that the relocated zirconium contributed to the formation of the crust layers and-was retained in-the central core region as metallic inclusions.
The high zirconium concentrations are present only in the metallic phase, and the L
data indicate that1the zirconium forming these layers did not participate in'the zirconium-water oxidation reaction and was transported as metallic zirconium or Zircaloy (melting point 2030 K) to form the layers in the central region of the ' core.
-This behavior suggests a core damage progression where the upper core region was not heated to high temperatures (>2200 K) until-after the upper crust had formed.-
1
-280-I
s The data for the debris relocated to the lower reactor vessel head
[*
indicate a lesser perMntage of zirconium (about 4% of core inventory) for 9
the quantity of uranium present.
These data suggest that some zirconium did not remain with the uranium relocated to the lower react vessel head, but contributed to the formation of the lower crust.
The summed data for the various repositories indicate that about 9% of the core inventorv ef zirconium has not been accounted for.
Although there is a relatively large uncertainty due to the few number of samples examined relative to the mass of the reactor core, these data suggest the presence of additional Zr repositories.
The tin distribution in the reactor vessel is significantly diffcrent than that observed for the Zr.
The core materials distribution data in Table 2 indicate that no significant amounts of Sn were found in the upper reactor plenum, the upper core debris bed, the debris on the lower reactor vessel head, and by inference from the lower reactor vessel debris data, in the lower and upper core support assemblies. The data indicate the presence of significant amounts (factors of 4 to 6 times greater than the core average concentratica - 0.3 wt%) in the upper and lower crusts and i
i the central core region. The data for the metallic region data indicate concentrations from 7 to 20 times the core average with higher concentrations in metallic inclusinns in the central core region.
The significant concentrations of Sn in the metallic region might be expected from the chemical behavior of Sn (i.e., a high free energy rquirement for oxidation),
in the lower crust of the central consolidated region, there is i
evidence of accumulations of Sn in the metallic samples; however, in contrast to the upper crust there are also accumulations of Sn in the ceramic samples relative to Zr. These data suggest a scenario where much of the Sn in the upper part of the core flowed down and was either trapped as metallic inclusions in the ceramic melt or contributed te formation of the metallic lower crust.
The inventory data for tin in Table 2 suggest that approx;mately one third of the structural tin originally present in the core is now located in the lower crust and that almost half is present in the central core region.
Control materials The THI-2 reactor cora contained 2200 kg of silver, 412 kg of indium, and 137 kg of cadmium control materials. These control materials were originally present in controls rods with 80% s'ilver,15% indium, and 5%
cadmium. During the accident at TMI-2, a significant fraction of this material was expected to melt and possibly volatilize due to the relatively low melting point of the alloy (1072 K - 1123 K), and the low boiling points of the individual constituents. Analyses indicate that at THI-2, control materials melted, and due to the pressure, dig 0 prohce a forcible ejection upon failure of the control rori cladding Table 3 lists the distribution of control material in the reactor vessel. The data indicate the presence of no measurable amounts of silver, the least volatile control material, on the lower reactor vessel
-281-
~ - _
e
ee r*
head, and by inference, in the upper and lower core support assembly 8
regicns.
The data also indicate that a small amount of silver (10%) was deposited on the upper plenum surfaces.
The bulk of this material was deposited with other fuel and structural materials as loose debris, which, based on particle size, could have been transoorted to the upper plenum as either a hydrosol or an aerosol.
Examinations of samples from outside the core indicate that only small amounts of the control materials were released from the core, substantiating tha analytical conclusion that a forcible ejection of the control materials probably did n:,t occur.
Below the upper plenum structure is a void and the loose debris bed, which accounts for approximately 20% of the core mass.
Based on the ar,alytical results, only 1.8% of the core inventory of silver is contained in this mass.
These data suggest that much of the silver (22%) melted and flowed down to form tn) crust layers and some was retained as inclusions in the central core region.
The silver data in Table 3 indicate that only about 47% of the core inventory of silver can be acceunted for. Although the uncertainties associated with these data are relatively large due to the small percentage of the whole core examined, the data suggest additional repositorieswhereasignificantfractionofthesilvergnventoryis located.
The high density of metallic silver (10.5 g/cm --near the highest of the core materials) would suggest that it is deposited near the bottom of the core or possibly on the lower head of the reactor vessel.
l This is consistent with the lower crust data where the silver concentration (4.5 wt%) is the highest of any of the core regions examined.
Table 3 indicates that indium is concentrated in the central core region at higher relative concentrations than was the silver. The highest concentrations are in the lower r. rust at 1.1 wt%, which is 3-4 times the core average.
This element is present in the metallic phases similar to silver.
However, a significant amount is also present in the cerainic parts at concentrations 2-3 times the core average.
If extrapolated to the mass of the various regions of the lower core, approximately 81% of the core inventory of indium is present in the lower core region.
Although this estimate has a relatively large uncertainty, the data indicate that approximately 30% more indium than silver is located in the lower core, which suggests a different distribution of indium than silver and indicates that a significant fraction of t M silver behaved differ ly than the indium and is located in another part of the reactor vessel The cadmium data indicate concentrations for the lower core that are similar or less than the cora average concentration of 0.1 wt%. The data indicate lesser concentra' % in the ceramic regions than the metallic regions, which might be exp.x.ted es cadmium is not expected to oxidize.
Much of the relatively volatile Cd was not released from the core but was 2
retained in the central molten part of the core as a gas during the early part of the accident, probably as inclusions in the molten metallic material.
-2 8'?-
WJ t-
A,.
e, TABLE 3. CONTROL R0D MATERIALS DISTRIBUTION IN THE REACTOR VESSEL Core material distribution Core material Percent of Inventerva Repository Silver Indium Cadmium Upper reactor plenum 1.0
-b-
-b-Upper core debris 1.8
-c-
-c-Upper crust region ceramic 1.2 3.6 0.65
- metallic 2.4 3.3 0.39 Consolidated region ceramic 10 27 6.1 metallic 1.6 2.'l 1.1 Lower crust region ceramic 7.3 7.2 1.4 metallic 11 16 29 Intact fuel rod >d gi 33 11 Lower reactor vessel head
-c-
-c-
-c-Lower core support asrembly
-c-
-c-
-c-Upper core support assembly
-c-
-c-
-c-l Total 47 70 23 a.
Percentage of the total amount of the element originally present in
.the core.
b.
Insignificant amount (<0.1 wt%) based on the upper plenum measurements.-
c.
Elemental constituent not detected based on detection-limits of approximately 0.1 wt%.
d.
Only-10.7% of the partially intact fuel assemblies contain control material as the balance (22.7%) are peripheral assemblies which do not -
contain control. materials.
-283-l J'