ML20040F125
| ML20040F125 | |
| Person / Time | |
|---|---|
| Site: | Calvert Cliffs  |
| Issue date: | 01/29/1982 |
| From: | ABB COMBUSTION ENGINEERING NUCLEAR FUEL (FORMERLY |
| To: | |
| Shared Package | |
| ML20040F123 | List: |
| References | |
| CEN-193(B)-NP, NUDOCS 8202080350 | |
| Download: ML20040F125 (53) | |
Text
- _ _ _ _ _ _ _
1 1
CEN-193(B)-NP l
l
=
N PARTIAL RESPONSE TO NRC QUESTIONS ON CEN-161 (B)-P, IMPROVEMENTS TO FUEL EVALUATION MODEL JANUARY 29, 1982 COMBUSTION ENGINEERING, INC.
8202080350 820202 PDR ADOCK 05000317 P
t 1
LEGAL NOTICE l
m
. g This report was prepared as an account of work sponsored by Combustion Engineering, Inc. Neither Combustion Engineering nor any person acting on its behalf:
Il A.
Makes any warranty or representation, express or implied including the warranties of fitness for a particular purpose or l
merchantability, with respect to accuracy, completeness, or useful-ness of the information contained in this report, or that the use of any information, apparatus, method, or process disclosed in this report may not infringe privately owned rights; or B.
Assumes any liabilities with respect to the use of, or for damages resulting from the use of, any information, apparatus, method or proces: disclosed in this report.
4 h
l l
l
g 1
l l
Question 8.
Fission Gas Release Data l
Question A.
The Bellamy and Rich data (Ref. 9) have been used to verify s
the fission gas release model. How were the rod linear ratings determined for these data as they were not given in the original publication? If this information is contained in Reference 9-2 of CEN-161, please provide us with a copy of this reference.
l Response A.
As stated in CEN-161(B) on p. 9-3, the fuel rod linear heat ratings were estimated to match the reported end-of-life centerline tegeratures, since the fuel temperatures and not the fuel heat ratings were reported in the original Bellany and Rich publication. The heat ratings were calculated for each rod by C-E using an iterative process which duplicated the reported boundary conditions of clad temperature (500 C) and peak fuel centerline temperature.
Fuel centerline tegeratures were reported for each rod according to assumed values of the gap resistivity, which was concluded to be in the 2
range of 1.5 to 2.0 cm - *C/w by Bellamy and Rich. The specific temperature profiles through the fuel pellets of each rod were determined in the C-E analysis by the use of the Lyon's inte-gral of thermal conductivity which, when evaluated between 0 C and 2800*C is 93 w/cm for 95". TD UO fuel. This conductivity was 2
modified for the 98". TD fuel using the standard Maxwell-Euken correction factor.
The fuel surface temperature is also related to the clad temperature by the expression (from footnote a, Tables 9-5 and 9-6 of CEN-161(B):
l
! L
r
^'
(
I Ts (fuel surface, 'C) = Tc (clad ID, *C) q 2
+ 10.44 x kw/ft x Gap Resistivity (cm
'C/w) 1 in Pellet Diameter (cm), where T = 500'C.
c
(_
Reference 9-2 of CEN-161(B) expanded the data of the original Bellary and Rich publication.
Centerline temperatures as a function of assumed gap conductance values were reported in this document for an additional eleven rods, for which only the gas release data had I
been reported in the original publication. As in the original publication, however, rod powers again were not included, thus these rods were analyzed in a similar fashion to the rods for which fuel temperature data were reported in the original publication. Corrected gas release values also were included in Reference 9-2 of CEN-161(B) for the rods 5050 and 5049..
b - -
t t
Question B.
Bellamy and Rich have estinated (Ref. 9) the gap resistivity of their 2
data to be between 1.5 and 2.0 cm -C/w.
In Tables 9-5 and 9-6 of CEN-161, Combustion Engineering has provided predictions for this 1
range of gap resistivity. There are, however, several high burnup 2
data included in Table 9-5 (2.0 cm -C/w, but these are missing 2
from Table 9-6 (1.5 cm -C/w). Why were these data omitted from this study?
Response B.
As stated in Response A, Reference 9-2 of CEN-161(B) expanded the data of the original Bellamy and Rich publication.
Unfortunately, for some of the rods, the detailed terperature 2
histories were not given for both 1.5 and 2.0 cm -C/w.
Thus calculations were made only for the single temperature histories reported for these rods.
l
> L
1 t
Question 10.
Annular Pellet Model Question A.
Does the inclusion of the annular model change the other fuel behavior models (e.g., relocation, swelling and fission gas release)? If not, why not?
If changes were made, what were they?
Response A.
The fuel behavioral models for annular fuel pellets are assumed to be the same as for solid fuel pellets.
Insufficient experimental data exist with which to justify any differences (although there are a number of annular test rods currently being irradiated which can eventually be used to verify behavioral models for annular fuel).
However, it is expected that differences are insignificant. Most are only indirectly dependent on pellet geometry. For example, fission gas release is a function of fuel temperature but the fuel temperature calculation is modified for annular fuel.
All modeling changes which have been made for annular fuel are described in CEN-161(B). These changes are the solution for the radial temperature distribution, the renormalization of the flux depression constants to ensure conservation of energy, the thernal expansion of the uncracked fuel region, and the inclusion of the central hole in gas void volume and pressure calculations.
, 3~-
1 g
1
. t Question B.
The equations used to calculate the temperature distribution in an annular fuel pellet areI provided in Section 5 of CEN-161.
However, the method used.to determine the flux depression constants A, B, and C is not given. Please explain how these constants are derived. Are the same constants used for both solid and annular fuel, or are these constants specifically derived for a given (solid or annular) design? We would expect that the use of constants derived for solid pellets would be slightly nonconservative for annular pellet designs.
Response B.
3 i
I The flux depression constants A, B, and C given in CENPD-139 for solid pellet designs are also used for annular pellet designs after renormalization for conservation of energy. Comparisons were made between the fuel pellet inner diameter temperatures calculated with constant A, B, and C with those using constant inverse
'I diffusion length K in the Robertson formulation (l)for varying pellet diameters. The diffusion length K and A, B, and C were selected such that the radial power distributions (and thus fuel temperature distributions) in the FATES 3 and Robertson
' formulations were identical for the solid pellet. Higher pellet inner diameter temperatures were calculated with the power distributions resulting from the Robertson formulation, but to a negligible degree.
For example, at a 13 kw/ft linear heat rate, the maximum,d,1fference was about 5 F.
These results indicate that it is suffihient to use solid fuel pellet flux depression constants for annular fuel pellets.
s (IIAECL 807 (CRFD-835), " Integral Kd9 in Fuel Irradiation",
J. A. L. Robertson, April,1959.
e
t Question 11. Fission Gas Production Model Question A.
The fission gas production model in FATES-3 uses a value of 30 cc/ mwd while the American Nuclear Society Working Group 5.4 (Ref.10) uses a value of 31 cc/ mwd. What data and references were used to develop the 30 cc/ mwd value?
l Response A.
t The fission gas production model used in FATES 3 is based on a production rate of 30 atoms of xenon and krypton per 100 fissions. At 200 Mev per fission this production rate is 30 cc/ MWD at STP. This basis is the same as that used II) by Kraftwerk Union AG (I) W. Hering and R. Manzel (Kraftwerk Union AG), "A New Fission Gas Release Model for LWR Fuel Rods. Comparison of Measured and Calculated Gas Release", Paper presented at the 80th Annual Meeting of the American Ceramic Society, Detroit, Michigan, (8-10 May 1978).
5.
i Question B.
When fuel thermal expansion or swelling is large enough to cause fuel-cladding contact, does the relocation model allow for relaxation of previous relocation? If not, it appears that unreasonably high interfacial pressures and gap conductance may result.
Response B.
Fuel pellet relocation accumulated prior to fuel-cladding contact is assumed to be irreversible. No additional relocation is calculated after contact. The interfacial pressure used in gap conductance calculations is always limited and gap conductance values are reasonable as indicated in Response A.
l I
D._
l 1
Question 12. Gap Conductance Model Question A.
The gap conductance model in FATES-3 no longer has an artificial limit on gap conductance, but calculates contact conductance based on fuel-cladding contact pressures. When fuel-cladding contact is made, what is the typical range of gap conductivity calculated?
Response A.
Contact conductance in the previous FATES code version was also based on the fuel-cladding contact pressure. Limits on gap conductance and fuel-cladding contact pressure were used. The fuel-cladding contact pressure limit is still used in FATES 3, but the limit on gap conductance has been removed.
The typical range in total gap conductances calculated 2
after pellet clad contact is 2500 to 3500 BTU /hr-ft,op, j
O ___ _
1 Question 13.
Helium Production Question A.
Postirradiation puncture data (Ref.11) from an experimental Halden fuel rod (Rod 8 of IFA-432) irradiated to 22,000 mwd /Mtu have shown more helium present (25%) than should be expected as a result of the initial fill gas int.'oduced during fuel rod fabrication. Similar behavior has been reported (Refs. 5-6) for the Riso rods mentioned previously. Does the FATES-3 code take into account helium production and release? If not, would the inclusion of such a model have a significant effect on calculated end-of-life rod pressures?
I Response A.
The FATES 3 code does not take into account helium production and release. The experimental rods cited (IFA-432 and RIS0) are pressurized to only one atmosphere with helium fill gas. Hence, the additional helium appears significant in these rods (i.e., 25% more helium at 22000 MWD /MTU in IFA-432).
However, such helium generation and release in prepressurized fuel rods would result in a partial pressure of released helium that would be expected to be an insignificant fraction of the partial pressure of the helium fill gas. Consequently, it would not have a significant impact on calculated end-of-life rod pressures.
In addition, as part of the EPRI/C-E sponsored fuel evaluation program at Calvert Cliffs 1, amounts of helium present in o
several fuel rods irradiated for one to four cycles were measured. The data for one, two, and three cycle fuel rods
-g-
1 have been published (I)'(2) and the data for four cycle fuel rods will be published in the near future (3)
Helium present in these rods before irradiation could be established from the work done on the archive fuel rods (1) as well as the detailed pre-irradiation characterization data available for these rods (4)
A review of these data shot that because of helium absorption early in life, amounts of helium collected after the first cycle are always lower than that initially present. Furthermore, very little additional changes in helium volume are observed after additional cycles of irradiation. These well-characterized data support the view that an insignificant amount of helium is released in pre-pressurized fuel rods.
(1) NSPD-75, " Gas Release and Microstructural Evaluation of One and Two-Cycle Fuel Rods from Calvert Cliffs I",
S. R. Pati, Combustion Engineering, Inc., Windsor, Connecticut, March 1979.
(2) C-E NPSD-119, " Gas Release and Microstructural Evaluation of Three-Cycle Fuel Rods from Calvert Cliffs I",
S. R. Pati, Combustion Engineering, Inc., Windsor, Connecticut, December 1980.
(3) " Gas Release and Microstructural Evaluations of Four-Cycle Fuel Rods from Calvert Cliffs I", S. R. Pati, Combustion Engineering, Inc., Windsor, Connecticut, to be issued to EPRI.
(4) " Fabrication and Characterization of BGAE-1 Test Fuel Assemblies",
V. Pasupathi, Combustion Engineering, Inc., Windsor, Connecticut, November 1975.
t ATTACHMENT 1 HIGH-BURNUP DATA ON FISSION GAS RELEASE FROM CALVERT CLIFFS I OBTAINED FOLLOWING THE SUBMITTAL OF CEN-161(B)
A.
Experimental Data As a part of a fuel performance surveillance program jointly sponsored by the Electric Power Research Institute (EPRI) and C-E (EPRI RP 586-1 Task A),
fission gas release data are being obtained from fuel rods irradiated in Calvert Cliffs I (CC-1) to various burnups. Data obtained from fuel rods irradiated through one to three cycles were reported in CEN-161(B). Analyses of the data using FATES 3 also were included in CEN-161(B).
Following the submittal of CEN-161(B), additional fission gas release measurements were obtained for six four-cycle rods, with rod-averaged burnups between 41.4 and 45.8 GWd/MTU, in a continuation of the Task A program. These results are reported in Reference 1 (attached), and also were previewed at l
an NRC/C-E meeting on safety related R&D, held in Windsor, Connecticut in December 1981 (Ref. 2). These new high-burnup data show that fission gas nelease from PWR fuel rods pressurized with pure helium, and containing either densifying or nondensifying fuels, !s low (less than 1%) and is essentially independent of burnup up to at least 46 GWd/MTU.
Within this group of six rods was a standard Batch B rod (NBD144) which contained densifying fuel, but was pressurized with 5% argon and 95% helium rather than '.vith pure helium, as were the five test fuel rods. The 1.5% gas release measured for NBD144 was slightly greater than the measurements for the test rods. This is believed to have resulted because the fuel in NBD144 had operated at higher temperatures through its entire irradiation history compared to the temperatures of the fuel in the rods containing pure helium.
A reduction in gap conductance in NBD144 due to the presence of argon is thought to be mainly responsible for the higher fuel temperatures. An additional factor was that NBD144 was a peripheral rod in the BT03 assembly and, therefore, had operated at somewhat higher heat ratings (especially in the third and fourth cycles) compared to other rods which were fabricated with fuels having the same enrichment, but were located in the interior of the assembly.
1-1
8 5
B.
Analysis by FATES 3 In CEN-161(B), fission gas release values predicted by FATES 3 for four CC-1 fuel rods containing the same densifying fuel type and irradiated through one to three cycles were compared against the measured values. This com-parison (in Fig. 2-10 o'f CEN-161(B)) showed that, for each of these four rods, fission gas release is conservatively predicted.
It is to be noted that densifying fuels similar to that used in Batch B in CC-1 are no longer used in the current product line of C-E.
Despite C-E's low interest in this old fuel type, three four-cycle fuel rods containing this densifying fuel (i.e.,
the two standard Batch B rods, NBD144 and
,, and Test Rod 09) were modelled in FATES 3.
In addition, one fuel rod containing a nondensifying fuel (i.e., Test Rod 54) was also modelled. The rod power histories used l
in the FATES 3 calculations for these rods included the local power ramping l
caused by changes in axial shape index (ASI), which occurred during Cycle 1 at CC-1 at a core-averaged burnup of 9.1 GWd/MTV. The change in ASI re-sulted during a xenon oscillation experiment described in Ref. 3.
(Note that this augmentation in rod local powers, which occurred for all of the fuel rods in Cycle 1, was not included in the histories of the CC-1 rods used in the gas release model developmental data base in CEN-161(B). This approach was a conservative one for model development.)
1-2
I i
l l
The FATES 3 predictions of fission gas release for the four-cycle rods mentioned above are compared against the measured values of gas release in Table 1.
Some key design and operating parameters for these rods are in-cluded in the table.
Furthermore, FATES 3 gives conservative predictions of the gas release measured for the other standard rod, NBD144, and the two test rods 09 and 54. These results provide further verification of the applicability of the FATES 3 code at extended burnup.
In the FATES 3 calculations described above, the densification model approved by the NRC for licensing calculations was used to describe the densification behavior of the fuels. Maximum density increases of 5.5 and 1.5% were used for the densifying and nondensifying fuels, respectively.
These values were derived from the post-irradiation density data obtained on these fuels as a function of burnup up to about 45 GWd/MTU (Ref. 4).
Following the procedure for licensing calculations, the fuels were assumed to reach their tenninal densities by 4 GWd/MTU, although the experimental data (Ref. 4) show densification continues to at least 20 GWd/MTU. To obtain an indication of the sensitivity of gas release predictions to the assumption of densification kinetics, another set of FATES 3 runs was made.
In the latter set, the densification kinetics in FATES 3 were changed to approximate the experimentally observed behavior. The same basic form of the equation as used in the licensing model was retained. Only the constants in the equation were changed so that the density increase of 5.5% occurred over 20 GWd/MTU.
The results of these calculations are included in Table 1.
It is apparent from the results that for all of the rods with densifying fuel, the conservatism of the predictions is reduced by the use of the alternate densification kinetics. No change in the licensing densification model is warranted, therefore, as this model provides the greater conservatism in application.
1-3
i REFERENCES 1.
S. R. Pati, H. R. Freeburn and L. V. Corsetti, " Fission Gas Release from PWR Fuel Rods under Conditions of Normal Operation ard Power Ramping,"
paper submitted for publication in Proceedings of ANS Topical Meeting on LWR Extended Burnup - Fuel Performance and Utilization, Williamsburg, Virginia, April 4-8, 1982.
2.
CENPD-267, " Safety Related Research and Development for C-E Reactors -
1931 Program Summary," (to be issued).
3.
J. Lippold et. al., " Field Experience with PWR Maneuvering Strategies Having Low Local Fuel Power Ramping," presented at Enlarged Halden Programme Group Meeting on Process Supervision and Control in Nuclear Power Plants, June 6-10, 1977, Fredrikstad, Norway.
4.
S. R. Pati and N. Fuhrman, "Densification Swelling and Microstructures
{
of LWR Fuels Through Extended Burnups," paper submitted for publication in ' Proceedings of ANS Topical Meeting on LWR Extended Burnup - Fuel Performance and Utilization, Williamsburg, Virginia, April 4-8, 1982.
i l
e i
1-4
TABLE 1 8
g COMPARISON OF MEASURED AND FATES 3 CALCULATED GAS RELEASE FOR FOUR-CYCLE CC-1 FUEL RODS FUEL R0D IDENTITIES FUEL PARAMETERS NBD144 09 54 standard Batch B test test Rod type Fuel-densification characteristic densifying densifying nondensifying i
Enrichment, % U-235 2.45 2.45 2.45 Initial density, % TD 93 93 95 Fill pressure, psig at'20 C 450 450 450 0
Fill gas composition' 5% Ar - 95% He 100% He 100% He Fuel-clTdding diametral gap, inch 0.0085 0.0085 0.0085 4.
OPERATING PARAMETERS No. of. Operating cycles 4-4 4
Rod-averaged burnup, GWd/MTU 42.2 41.4 41.4 Location within the assembly periphery interior interior Local Peak Power at 9.1 GWd/MTU, kw/ft 8.6 8.6 8.6 Measured E0L Gas Release, %
'1.45 0.36 0.55 FATES 3 Predicted Gas Release,,%
11.4 10.0 1.0 FATES 3 Predicted Gas Release'for,
8.2 6.4 Modified Densification Kinetics, %
The FATES 3. power histories for these.four-cycle rods include a local power augmentation in the first cycle due to a xenon oscillation experiment. This power augmentation was not included in the. power history of CC-1 fuel rods analyzed by FATES 3 in CEN-161(B).
- Maximum density increase is assumed to take place over 20 GWd/MTU, rather than 4 GWd/MTU.
8 s
FISSION GAS RELEASE FROM PWR FUEL RODS UNDER CONDITIONS OF NORMAL OPERATION AND POWER RAMPING S. R. Pati H. R. Freeburn L. V. Corsetti Combustion Engineering, Inc.
(
Nuclear Power Systems Windsor, Connecticut Measurements of fission gas release are presented for helium-pressurized fuel rods of PWR design which were all fabricated using Zir-caloy-4 tubing as the clad material and UO2 fuel in solid pellet form.
Various fuel microstructures are included which have distinct differences in grain size, density, porosity distribution and propensity for densification in reactor. One group of gas release measurements is from fuel rods irradiated through one-to-four cycles at Calvert Cliffs I.
These measurements show that less than one percent fission gas release can be expected from the fuel pellets in modern PWR fuel rods for normal operating conditions in power reactors and to burnups at least as high as 46 GWd/mtU, with no strong enhancement of gas release fraction with burnup. Furthermore, the microstructural differences which existed in the as-fabricated state did not markedly affect the amount of gas re-leased among the fuel types examined. The results presented for the Calvert Cliffs I rods are further corroborated by comparison with measure-ments of gas release from other PWR fuel rods up to burnups of about 40 GWd/mtU. This work was sponsored jointly by the Electric Power Research Institute and Combustion Engineering, Inc.
Another group of gas release measurements is presented for pres-l surized PWR fuel rods which have undergone fast power ramps in a test re-l actor at Studsvik, as part of an internationally sponsored program, following up to three cycles of power reactor irradiation. The peak linear heat ratings achieved during ramping are well above the normal heat ratings typical of PWR operation. In contrast to the Calvert Cliffs I results, the measurements of gas release for the ramped rodlets:
(1) are in many cases significantly greater than one percent; and (2) show strong depen-dencies on peak linear heat rating and on the grain size of the UO. Further-2 more, some evidence of a burnup enhancement in this high fuel temperature regime of operation is apparent among the measurements, but only through a burnup of about 25 GWd/mtU. This enhancement at high heat ratings is thought to be strongly influenced by changes in gap conductance which tend to stabilize at higher burnups due to gap closure effects.
8 i
FISSION GAS RELEASE FROM PWR FUEL RODS UNDER CONDITIONS OF NORMAL OPERATION AND POWER RAMPING
~
INTRODUCTION The possibility of a large increase in fission gas release from UO2 fuel pellets is a factor being considered when targeting fuel rods in an LWR for l
extended burnup. The collection of well-characterized experimental data on fission gas release under' normal and off-normal conditions has been an important contribution to the evaluation of the effects of fission gas release on the overall performance of fuel rods at both standard and extended burnups. Under normal operation, the heat ratings of fuel rods in a PWR generally decrease with time. The heat ratings toward the targeted discharge burnup are signifi-cantly below the heat ratings that are associated with licensing calculations.
Therefore, it may seem that for licensing calculations, high burnup gas re-lease data obtained under the conditions of normc1 operation are of secundary
(
importance. However, at the heat ratings associated with safety analyses, and at burnups above 25-30 GWd/mtU, more than 50% of the volume of fuel in an LWR fuel rod remains at temperatures which are in the lower temperature range.
The data from normal operation, cherefore, are very valuable in developing temperature and burnup dependencies of fission gas release over an important range of these variables.
The value of gas release data obtained from fuel rods ramped to higher powers, following their base irradiations in power reactors, cannot be over-emphasized. Heat ratings during base irradiation are generally low, and as a result, essentially all of the generated inventory of fission gas is stored in the fuel. This stored inventory becomes available for release during any sub-sequent ramping to higher powers. This type of history promotes higher gas release than that generally observed from fuel rods which experience higher powers early in life, followed by power decreases with time.
A limited number of gas release data on fuel rods irradiated under normal operating conditions of LWRs to burnups beyond 25 GWd/mtU has recently become available 1-5 However, well-characterized data at burnups above 35 GWd/mtU for normal conditions and at moderate to higher burnups for off-normal condi-tions have been sparse. Such data are presented here and apply to modern PWR fuel rod designs.
FISSION GAS RELEASE DURING NORMAL OPERATION IN POWER REACTORS EPRI/C-E PROGRAM AT CALVERT CLIFFS I The Electric Power Research Institute (EPRI) and Combustion Engineering (C-E) have been conducting a joint fuel performance evaluation program in Calvert Cliffs I (CC-I) since 1975. A major objective of this program is to
8 s
evaluate the performance of test rods'through five operating cycles. These test rods have systematic design variations in microstructure, density, pro-pensity for in-reactor densification and rod internal pressure. Design and fabrication details for the fuel rods included in this test program have been published by Pasupathis, A unique feature of the EPRI/C-E program at CC-I is that it allows the effects of different design variables, including fuel type, to be evaluated in well-characterized test rods irradiated under nearly identical operating con-ditions in a power reactor. Consequently, performance comparisons among the fuel types can be made without the uncertainties attached to different operating conditions and irradiation environments. This is an important consideration for any experiment if it is to provide high quality fission gas release data that is suitable for modeling purposes or mechanistic evaluations.
Thus far, test fuel rods have been irradiated for four operating cycles at CC-I and have received detailed examinations at poolside during each of the refueling outages. The results of these examinations have been reported by Bessette et al.,7 and Ruzauskas et al.,8-9 After each cycle, a number of the test rods were selected for additional examination at a hot-cell facility. The specific objectives of the hot-cell examinations were: to measure fission gas release from fuel types typical of those currently being used in commercial PWR systems; to obtain high burnup gas release data from well-characterized fuels for evaluating possible burnup or thermal dependencies and to evaluate the effects of different fuel micro-structures and other rod design variables on gas release. The fission gas release data obtained after Cycles 1 and 2 with rod-averaged burnups to 29.1 GWd/mtU have been discussed previously by Pati, et al.,1 Recently obtained fission gas release measurements from rods after three and four cycles of irradiation (average burnups of up to 45.8 GWd/mtU).will be discussed in this paper together with the previously obtained data to evaluate possible dependencies on temperature and burnup.
FUEL ROD SELECTION AND HOT-CELL PROGRAM WORKSCOPE The rods selected for the hot-cell examination were chosen so that gas release could be investigated as a function of the following:
Power and Temperature - Rods with different initial enrichments were selected. Temperature variations also resulted from the use of densi-fying and non-densifying fuels of the same enrichment and initial density.
One rod containing densifying fuel had a different initial fill gas composition based on a 95% helium - 5% argon mixture instead of pure helium.
Burnup - Rods with essentially identical design variables were examined after each operating cycle.
Fuel Type and Microstructure - Non-densifying and densifying. fuels of similar enrichment were included. There were also microstructural differences among the four types of non-densifying fuels examined.
l The 18 test fuel rods selected for gas release measurements in this test l
program are listed in Table 1.
The calculated rod-averaged burnups range I
i L
s from 18.7 GWd/mtU af ter one cycle to 45.8 GWd/mtU af ter four cycles.
Other than pellet geometry, initial fill pressure and fill-gas composition, the i
designs of the rods are similar. These particular rods were selected from d
.; hose in the overall irradiation experiment such that differences in power histories were minimized in rods with the same initial enrichment. Four non-densifying fuel types and one densifying fuel type representative of the stan-I dard Batch B fuel in CC-I were included in the rods sent to the hot cell.
]
It may be noted that the densities and grain sizes of the non-densifying fuel types were different. These variations were introduced by using different pore-f 6
formers and fabrication histories. For :::..gle, the Type III fuel was fab-I ricated with HMA poreformer and had an initial density of 93% TD with a grain size of 7 pm.
By comparison, the Type II fuel, fabricated with PVA poreformer, had an initial density of 95% TD and a grain size of 4 pm.
The higher enriched (Type 7) non-densifying fuel was also fabricated with PVA poreformer, but had
}
a larger grain size (15 pm) and lower initial density (93% TD) than the Type II fuel fabricated with the same poreformer.
There were also differences in porosity distribution among the non-densifying fuel types because of differences in the type and amount of pore-formers used and differences in sintering conditions. For example, about 70%
of the total porosity in the Type V fuel (93% TD, PVA poreformer) was open with the pores concentrated at coarse agglomerate boundaries. Approximately 45% of the total porosity exhibited this feature in the Type II fuel (95% TD, PVA poreformer). The percentages of the total porosity which were open for fuel Types II and IV vere 60% and 10%, respectively. In contrast, essentially all of the porosity in the densifying fuel (Type I) was intra-agglomerate and 3
closed.
Several important details of the operating history for each of the rods examined for gas release are summarized in Table 1.
Differences in operating power among the various fuel types were achieved by using different enrichment levels. The peak local linear heat rating for all fuel types occurred early in the first operating cycle, as shown in Table 1, with the maximum being approxi-235 ) Type V, non-densifying mately 11.1 ku/ft for the high enrichment (2.82 wt%
U fuel. The enrichment differences among Types II, III and IV (i.e., 2.45 wt%
235 ) resulted in only small power differences.
235U vs. 2.33 wt%
U The workscope of the hot-cell program included fuel rod puncturing, gas collection and mass-spectrometric gas analyses. Other measurements performed included: gamma-scanning; end-of-life internal void volume and internal pressure; Details of the hot-cell examina-fuelburnupanalyses;andfuelmetallogaghy. The fission gas release results and tions were summarized previously by Pati the role of pertinent variables affecting these results will be discussed in this paper.
FISSION GAS RELEASE AND RELATIONSHIP TO OPERATING HISTORY AND FUEL MICROSTRUCTURE Data on the fractional fission gas release for the test fuel rods from CC-I are summarized in Table 1 and are plotted as a function of burnup in Figure 1.
Putting aside the data point from the single densifying fuel rod containing argon for the moment, a review of these data indicates that the gas release of the 3-and 4-cycle rods was low (< 1%), regardless of differences in fuel types. This is consistent with the behavior observed previously in the
i 1-and 2-cycle rods. Also, the fractional fission gas release does not exhibit an appreciable burnup dependence up to 45.8 GWd/mtU. Over the range of burnup thus far, slightly more gas release is observed in rods containing fuel Type V, which had the higher enrichment. This difference is consistent with the higher heat ratings and the greater as-fabricated open porosity of the fuel used in these rods. This aspect is discussed in more detail later in this paper, Peak and average heat ratings of the 18 fuel rods examined in this pro-s gram are listed in Table 1.
Detailed physics data on power history, axial power shape and burnup distributions for the rods are presented in references 7-9.
To better understand the role of temperature history and the fuel micro-structure, fuel temperatures of a select group of rods were calculated for several burnup periods. A coolant temperature of 570F (300C) was assumed for 10 all fuel temperature calculations. The Lyons integral of thermal conductivity for 95% TD UO2 (i.e., 93 w/cm between OC and 2800C) was used to develop the i
radial temperature profile through the CC-I fuel pellet. If different from 95% TD, this conductivity was modified using the Maxwell-Eucken correction for density described by Hann, et a111 The change in fuel bulk density due to densification and swelling effects was also accounted for.
In-reactor densi-fication and swelling behavior of the fuels associated with these rods were known from the post-irradiation density measurements of a number of samples over the burnup range 9-50 GWd/mtU12 These density data, post-irradiation 2
fuel-clad gaps observed on metallographic samples,3, as well as the cladding profilometry data ~9 were used as inputs in estimating the relative differences 7
in gap conductance for the fuel rods. To provide a condensed version of the temperature history, the temperature calculations were performed at the following burnup periods:
(1) at the time of peak heat rating early in life with the assumption that the fuel attained its terminal density resulting in the largest gap and thereby providing an indication of the peak fuel temperature through life; and (2) at the approximate mid-point of each cycle of exposure using the axial and time averaged heat rating for the cycle to obtain an estimate of the cycle-averaged fuel temperature. A modest correction to the radial temperature profile through the fuel pellet was included to account for flux depression effects in the CC-I fuel. The gap conductance assumptions and the results of the fuel centerline temperature calculations are shown in Table 2.
An inspection of the calculated temperature histories of the CC-1 fuel rods listed in Table 2 indicates that, except for very early in life, the fuel temperatures were low. As discussed later, the low temperature in this context refers to the fact that the contribution to fission gas release from processes other than knockout and recoil were minimal in this temperature range. This factor was primarily responsible for the low gas release observed in all of these rods.
Among these low release values, however, small differences exist.
A correlation appears to exist between the individual release values, the fuel types and the associated temperature histories. After each cycle of irradiation, the fuel rods loaded with fuel pellets having 2.82 wt% 235U released more gas than the rest of the fuel rods pressurized with helium. As indicated by the temperature history in Table 2, the peak fuel centerline temperature reacned 1490C in these fuel rods early in the first cycle of irradiation. Metallography showed an accumulation of porosity occurred at grain boundary triple points at 2
the center which extended to approximately 10% of the pellet radius. This
i observation is consistent with the above temperature history for fuel of this j
microstructure and level of open porosity. The high operating temperature in j
these fuel rods, occurring early in life, may have been partly responsible for the higher release observed after the first cycle of irradiation. However, the trend of a slightly higher level of release was maintained through the four cycles of irradiation, although during this period, the operating temperatures of the different fuel types were approaching each other because of nearly J
identical heat ratings and comparable gap conductances late in life. The s
continued differences in gas release behavior through four cycles appear to be related to different levels of open porosity in the fuel. The fuel enriched with 2.82 wt% 235U had the highest as-fabricated level of open porosity (about 70% of the total porosity was open) among the various fuels. A higher level of open porosity generally leads to a higher release due to direct fission product ktsckout with the higher surface-to-volume ratio in the fuel. Most of the fuel, even in the fuel of higher enrichment, operated at temperatures where knockout of fission gas was the predominant mode of gas release. The dominance of the knockout mode in this temperature range (< 800C) is also apparent i
from the observation that this fuel released two to three times more gas than the densifying fuel despite the fact that its graf r tize was six times larger j
than the lattar fuel type (15 vs. 2.5 pm).
If the diffusion of fission gas l
was playing a prominent role in controlling gas release, the larger grain size fuel would have shown lower fractional fission gas release than the fuel of smaller grain size. Thus, the insensitivity of gas release to fuel grain size is an indication that the fuel temperatures for most of the irradiation period were in the range where gas release was essentially independent of temperature.
Similarly, in the case of the nondensifying fuel Types II and III after three cycles of irradiation, the latter fuel type showed almost twice the fractional fission gas release of the former fuel type. In contrast, the Type III fuel had a higher as-fabricated grain size (7 pm) than the Type II (4 pm) fuel. Higher gas release in the Type III fuel is believed to hsve resulted primarily from a higher level of open porosity compared to the Type II fuel (2.2% vs. 1.3% of the pellet volume).
Almost identical gas release was measured in the two fuel types with small grain size, Types I and II, irrespective of differences in densification stability. As shown in Table 2, the greater propensity for densification in the Type I fuel led to substantially higher temperatures than experienced by the Type II fuel in the early part of its first cycle. Because of the above, only the densifying fuel showed a limited, but measurable, amount of grain growth. The grain growth was observed at the axial position corresponding to tion,3.y-in-life peak power locations, after two and three cycles of irradia-the earl 2
The results suggest that the grain growth occurred in this fuel during the early part of irradiation. Furthermore, no change in grain size occurred with continued irradiation to higher burnups because the fuel temperatures decreased with burnup. Despite the differences in thermal history, the EOL gas release values were almost identical for the following reasons:
(1) the inventory of fission gas was low during the period when the temperature difference between the two fuel types was greatest and (2) the densifying fuel had a very low level of open porosity compared to the other fuel type. The low open porosity counterbalanced the effect on the gas release caused by the continued operation at higher temperatures during most of the first cycle.
I s
4 The higher gas release r.easured for Rod NBD144, which contained 5% argon mixed with helium, resulted from a higher temperature of operation through the entire irradiation history compared to the temperatures of sibling fuel rods containing helium. A reduction in gap conductance due to the presence of argon was mainly responsible for the higher temperatures.
In addition, Rod NBD144 was a peripheral rod in the assembly and operated at somewhat higher heat ratings (especially in the third and fourth cycles) compared to other rods fabricated with fuels of the same enrichment, but located in the interior of the assembly (e.g., Rod 09).
COMPARISON WITH OTHER DATA The fission gas release data from CC-I have been plotted in Figure 2 along with data obtained from fuel rods irradiated in four other PWRs, namely, H.B. Robinson 13, Point Beach-ll4, Oconee-215, and Zorita16 As in the case of the CC-I data, small differences in individual release values observed within a given set of fuel rods could possibly be explained by differences in operating histories and fuel microstructures among the rods. For example, a value of 0.9% release measured in a Point Beach-1 fuel rod at 20.9 GWd/mtU was higher than in the rest of the Point Beach-1 fuel rods. This rod experienced a larger increase in heat rating at the beginning of Cycle 3 over the average of Cycle 2 compared to the others in the set.
Similarly, higher gas release observed in some of the Zorita fuel rods at 40 GWd/mtU compared to the other PWR data may be related to differences in operating history between these Zorita test rods and the typical commercial PWR rods. For example, the linear heat ratings of the Zorita test rods in the second cycle (6.6 kw/ft) and third cycle (6.2 kw/ft) were significantly higher than in the first cycle (4.6 kw/ft). This type of power history is not typical of the PWR fuel rods included in Figure 2.
Con-sequently, even in the low gas release regime, the time varying fuel temperature history is important for accurately evaluating gas release behavior. Thus, the combined data from the PWR rods in Figure 2 indicate that, under normal duty cycles typical of commercial PWR operation, the fission gas release fraction is low and essentially independent of burnup up to at least 45.8 GWd/mtU.
The conclusion that burnup plays a minor role in controlling fission gas release behavior in the normal operating range of LWRs has also been arrived l7 at by Ocken and Roberts This resulted from their review of data obtained under EPRI sponsorship from fuel rods irradiated in seven LWRs. The data base included fission gas release data of one-and two-cycle CC-I fuel rods generated under this program. Based on their review, the authors concluded that fuel temperature is the key variable which determines fission gas release.
At the time of that review, the data were available up to a maximum burnup of 29 GWd/mtU. The newer data presented and analyzed herein extend the validity of the above conclusion to a burnup of up to 446 GWd/mtU.
THERMAL VS. BURNUP DEPENDENCE OF GAS RELEASE BELOW 1250C Gas release at low temperatures was investigated by Bellamy and Rich 18, who measured gas release in the range of 0.1 to 0.2% up to 20 GWd/mtU in fuel of 98% TD with centerline temperatures below 1250C. A sharp increase in gas release (to 42.6%) was observed at burnups of about 35 GWd/mtU. The trend of increasing gas release continued up to 48 GWd/mtU. The sharply higher release observed at 35 GWd/mtU compared to that at 20 GWd/mtU was attributed
=. _..
i I
t i
to an acceleration in the rate of release due to knockout and recoil. The enhancement of gas release was interpreted as due to an interconn.ction of fission gas bubbles along grain boundaries which significantly increased the surface-to-volume ratio of the fuel. Although no data were available between j
20 and 35 GWd/mtU, the gas release acceleration was generally believed 18-20 to
{
occur beginning at burnups of 420 GWd/mtU.
The gas release data of Bellamy and Rich associated with fuel that i
1 operated below 1250C are plotted in Figure 2 to highlight the difference between these data and the previously mentioned PWR data. The difference between j
the two groups of Cata may be explain.d by differences in temperature history and the presence of several gas release mechanisms (knockout, recoil and thermal di. fusion) operating at fuel temperatures below 1250C. Below 1250C, no temperature dependence on gas release was considered by Bellamy and Rich i
i since the knockout of fission gas atoms was assumed to be the predominant l
mechanism. Thus, all of the gas release data at fuel centerline temperatures
[
up to 1250C were treated as one group and were considered to be independent i
i of temperature. However, some of the more recent fission gas release models show that diffusion effects may become significant at fuel temperatures below 1250C, especially at extended burnups. For example, calculations by Hargreaves and Collins 21 show that at 1200C a sharp increase in gas release due to the
{
diffusion-induced accumulation of gas atoms at grain boundaries occurs at about r
1 17 GWd/mtU. Thus, the increase in gas release observed by Bellamy and Rich is believed to have, occurred due to a contribution from diffusion effects.
i This explanation is consistent with the calculations of Bellamy and Rich which show that the specific surface area of the high-burnup specimens increased by a factor of about 25 over their low-burnup companions.
On the other hand, for the CC-I and other PWR rods, the low level of gas release continued up to 45.8 GWd/mtU. The factor responsible for these low releases was the continued operation at low operating temperatures whica mini-mized the diffusion effects. In the peak power region of some of the CC-I fuel rods, ti fuel temperature was high enough so that diffusion effects, such as j
grain growth and the formation of porosity at grain boundary triple points were j
noticeable in small fuel volumes. However, these occurred early in the first cycle when the fission gas inventory was low and, therefore, the gas released l
was minimal. In contrast, the thermal history of fuel rods reported by Bellamy and Rich 18 suggests that the high burnup specimens were subjected to an appreciable temperature excursion (still below 1250C) in their last cycle of irradiation, when the fission gas inventory was signif! cant. In addition, the 400C cladding temperature of the Bellamy and Rich fuel rods was signifi-cantly higher than the typical cladding temperature of normal PWR rods. This
[
suggests that for a given fuel centerline temperature, diffusion effects would be operative in a larger fuel volume fraction in the Bellamy and Rich rods than in typical PWR rods. The explanations offered above suggest that thermal 1
effects (as opposed to burnup effects) are mainly responsible for the higher gas i
release observed in the high-burnup fuel of Bellamy and Rich.
i DATA FROM POWER RAMP TESTS Hollowell, et al., have reported on the results of the Over-Ramp Program 22 at Studsvik in Sweden Only the variables considered pertinent to fission gas release are addressed in this paper. As part of the workscope in that program, i
8 i
fission gas release was measured in a number of pressurized PWR rods having UO2 solid pellets ramped subsequent to their base irradiation in power reactors.
The fuel rods were subjected to fast power ramps (%3 kw/ft-min) in the R-2 reactor after base irradiations in the Obrighaim (KWO) or BR-3 reactors.
The fuel rod designs tested varied from short segmented rods (about 40 cm long) irradiated in KWO to rods of longer lengths (about 100 cm long) irradiated in BR-3.
The rod-averaged heat ratings during the base irradiations ranged from 4.6 to 7.9 kw/ft, and the rod-averaged burnup levels ranged from 12 to 31 GWd/mtU. Fuels of four different grain sizes (4.5, 6,10.5 and 22 pm) were tested. Figure 3 is a plot of the gas release measured in the ramped rods, againrt the ramp terminal power level, at the peak power position in the rods.
A series of straight lines represents points of a common design, at a given level of burnup. The initial grain size of the fuel is shown on each set of data.
It is apparent from Figure 3 that fission gas release is low at ramp terminal power levels below 10.5 kw/ft for all fuel types. This observation is l
consistent with the behavior observed for commercial rods which have experienced normal irradiation in power reactors. A review of the power histories of these rods during the base irradiations has indicated that the gas released during the base irradiation is expected to be small relative to the release measured after the power ramps at R-2.
Since a significant part of the fuel column in the longer fuel rods ex-perienced local ramp terminal powers below 10.4 kw/ft, the fission gas release shown in Figure 3 for these rods has been adjusted. Specifically, this adjust-ment ignores the fission gas inventory for that portion of the fuel column below 10.4 kw/ft in the determination of the percentage of fission gas released during the ramp. Therefore, all of the data points in Figure 3 represent the release of fission gas from fuel ramped to local terminal powers above this linear heat rate.
Figure 3 demonstrates that the main variables which affect fission gas release are rod power (fuel temperature), fuel burnup and fuel grain size. For a given fuel type and burnup, fission gas release is strongly dependent on power (fuel temperature). A burnup dependence of gas release is evident by comparing release values of 6 pm grain size fuel at two reported levels of burnup.
In the range of ramp terminal powers of 13.4 to 14.9 kw/ft, fuel preirradiated to higher burnup (%24 GWd/mtU) releases more gas on a percentage basis than fuel at the lower burnup (413 GWd/mtU). Since the inventory of generated fission gas increases with burnup, increasingly more fission gas atoms will be released to degrcde the gap conductance, thus, contributing to higher fuel temperatures and larger percentage releases. This is in addition to any burnap ichancement occurring between these burnups.
The above trend of increased percentage release between 12 and 24 GWd/mtU appears to be reversed on increasing the burnup beyond 24 GWd/mtU. For example, neglecting the small difference in grain size between 4.5 and 6 um, a small reduction in gas release on a percentage basis is indicated when the release values observed at %24 GWd/mtU are extrapolated to lower powers at which data are available on loel with N30 GWd/mtU. An improved gap conductance with in-creasing burnup beyond the onset of fuel-cladding contact may be considered as a factor affecting the apparent burnup dependence in the above sets of data. The
I t
improvement in gap conductance may occur due to higher contact pressure at the fuel cladding interface which outweighs the degradation effect of increased gas release on a total-atoms-released basis. Therefore, it is possible to hypothesize that, at a given power level, the fuel temperature is reduced sufficiently at %30 GWd/mcU compared to fuel temperatures at s24 GWd/mtU such that any detrimental burnup effect is overcome by a beneficial effect of lower temperatures. Additional data at higher ramp terminal powers and for fuels with 30 GWd/mtU burnup and beyond are needed for a better understanding of these interrelated factors.
The pronounced effect of grain size is apparent from a comparison of the release values of different fuel types having a common level of burnup.
For example, at a burnup level of approximately 25 GWd/mtU, and at a ramp ter-minal power level of 13.7 kw/ft, the fuel with 22 pm grain size shows a factor of six lower gas release compared' to the 6 pm grain-size fuel. The data from 0
the fuel with an intermediate grain size of 10.5 pm follow the same trend.
Despite its significantly lower burnup at identical ramp termiani powers, the fuel with 10.5 pm grain size released two to three times more gas than the fuel with 22 pm grain size.
COMPARISON OF FISSION CAS RELEASE BEHAVIOR UNDER NORMAL IRRADIATION AND POWER RAMPING A distinct feature of the data from the ramp tests is that the gas re-lease is strongly dependent on temperature, in contrast to the essentially temperature independent (< 800C) nature of gas release under conditions of nor-mal irradiation. At powers above 11 kw/ft for pressurized PWR fuel rods, for which a significant amount of additional release takes place, the fuel tem-peratures are high enough for diffusion effects of fission gas to play a demi-nant role.
To obtain an indication of the temperatures at which such effects become apparent, a parameter study on fuel temperature was made. Fuel temperatures were calculated with four different assumed gap conductance values of 0.5, 2
1.0, 1.5 and 2.0 w/cm -C.
The coolant temperature was assumed to be 590F (310C). The Lyons integral of thermal conductivity for 95% TD UO2 (i.e., 93 w/cm between Oc and 2800C)10 was used to develop the radial temperature profile through the fuel pellet. Uniform radial distributions of material density and power density also were assumed for the calculation of the radial temperature profile. The results of the parametric study are presented in Figure 4.
This figure shows that the diffusion effects on gas release become dominant when '.he l
fuel centerline temperature exceeds 1300C, which will occur somewhere between 9-11 kw/ft, depending on the actual value of the operative gsp conductance. At higher temperatures the sweeping by grain boundaries may add to the release of fission gases. Since the diffusion effect is a combined function of time and temperature, the threshold temperature for higher gas release controlled by diffusion is expected to go down with increasing burnup accumulated during base irradiation. Thus, the results of these calculations are consistent with the l
interpretation provided earlier for the Bellamy and Rich dat: 1.
Diffusion i
affected the high burnup gas release behavior in the Bellamy and Rich data at j
fuel centerline temperatures below 1250C because the associated burnups i
were higher than 35 GWd/mtU.
8 s
Another distinct feature of the gas release data from ramp tests is that gas release is strongly dependent on grain size. In contrast, under normal irradiation, no grain size effect was observed. This difference in behavior can be explained by the operation of different controlling mechanisms of gas release in these two sets of' data. Under normal operation the fuel tem-perature during most of the irradiation is low so that the gas release is essen-tially independent of temperature and controlled by knockout and recoil. As a result, no sensitivity to grain size is observed. At the temperatures associated s
with the Studsvik ramp tests, gas release is controlled by diffusion of gas atoms and fission gas bubbles to the grain boundaries and by the sweeping effect of grain boundary movement during the grain growth process.
CONCLUSIONS An evaluation of the combined data from power reactor and ramp test f
programs has led to the following conclusions:
i l
o During normal operation typical of PWRs, fractional fission gas release j
from fuel rods pressurized with helium is low (generally less than 1%) and is essentially independent of burnup up to at least 46 GWd/mtU.
o The normal operating temperatures of the fuel in rods which have been pressurized with helium are sufficiently low at high burnups (centerline temperatures in the range of 600-800C) that the fuel microstructure, the fuel densification behavior, and the time-varying thermal history do not markedly affect fission gas release as measured at the end of life.
In this regime of temperature, fission gas release is essentially independent of temperature and is controlled by knockout and recoil mechanisms.
o Within these small releases, however, differences exist in the data which can be related to:
fuel microstructure (e.g., open porosity); composition of fill gas, especially when other than pure helium; and details of the fuel ther-mal history.
o Uncertainties in gap conductance preclude a precise estimation of fuel temperatures above which diffusion-controlled mechanisms of gas release become dominant. A comparison of the steady-state PWR data with the Bellamy and Rich test data suggests, however, that at burnups above 35 GWd/mtU, effects of diffu-sion-controlled gas release mechanisms become measurable at temperatures above 800C, which is below the previously suggested value of 1250C.
o At linear heat ratings exceeding 12 kw/ft, the fission gas release from pressurized PWR rods is much greater than 1%, is strongly dependent on power (temperature) and fuel grain size, and displays an apparent burnup enhancement to at least 25 GWd/mtU. More ramp test data above 12 kw/ft are needed, however, to determine the nature of any burnup dependence above 25 GWd/mtU.
Industry programs are now in place to obtain these data.
s ACKNOWLEDGEMENTS The collection and evaluation of data from Calvert Cliffs I fuel rods was performed under the joint sponsorship of the Electric Power Research ~
Institute (EPRI Research Project 586-1 Task A) and Combustion Engineering, Inc.
The assistance of Baltimore Gas and Electric's Calvert Cliffs personnel and Battelle Memorial Institute's Hot-Cell Facility personnel on this task is gratefully acknowledged. Our appreciation is extended to the participants of the Studsvik Over-Ramp Project for the use herein of the fission gas release data generated by that project. The authors would also like to express gratitude to their colleagues at C-E who have helped them in various phases of these multi-year tasks.
REFERENCES I
S. R. Pati, et al., " Fission Gas Release and Dimensional Changes of Test Fuel Rods Containing Densifying and Non-Densifying Fuel," Proceedings of ANS Topical meeting on LWR Fuel Performance, April 29 - May 3, 1979, Port-land, Oregon, p 303.
S. R. Pati, " Gas Release and Microstructural Evaluation of One-and Two-
[
2 Cycle Fuel Rods from Calvert Cliffs I," NPSD-75, Combustion Engineering, Inc., Windsor, Connecticut, March 1979.
3 S. R. Pati, " Gas Release and Microstructural Evaluation of Three-Cycle Fuel Rods from Calvert Cliffs I," C-E NPSD-119, Combustion Engineering, Inc., Windsor, Connecticut, December 1980.
4 E. Roberts, et al., " Westinghouse Pressurized Water Reactor Fuel Develop-ment and Performance," Nuclear Energy, 19 No. 5 (1980), p 335.
5 L. F. A. Raven, "Effect of Pressurization on the Release of Fission Cas from UO2 at High Burnup," Presented at the 82nd Annual American Ceramic Society Meeting, Chicago, Illinois, April 1980.
6 V. Pasupathi, " Fabrication and Characterization of BG&E-1 Test Fuel Assemblies," Combustion Engineering, Inc., Windsor, Connecticut, November 1975.
7 D. E. Bessette, et al., " Examination of Calvert Cliffs I Test Fuel Assemblies at End of Cycles 1 and 2," NPSD-72, Combustion Engineering, Inc., Windsor, Connecticut, September 1978.
8 E. J. Ruzauskas, et al., " Examination of Calvert Cliffs I Test Fuel Assembly after Cycle 3," NPSD-87, Combustion Engineering, Inc., Windsor, Connecticut, September 1979.
9 E. J. Ruzauskas, et al., " Examination of Calvert Cliffs I Test Fuel Assembly Af ter Cycle 4,"
C-E NPSD-146, Combustion Engineering, Inc., Windsor, Connecticut, October 1981.
8 s
10 M. F. Lyons, et al., " Power Reactor High Performance UO2 Program,"
GEAP 5591, General Electric Company, San Jose, California, 1968.
11 C. R. Hann, et al., "GAPCON-THERMAL-2: A Computer Program for Calculating the Thermal Behavior of an Oxide Fuel Rod," BNWL-1898, Battelle Pacific Northwest Laboratories, Richland, Washington, November 1979.
s 12 S. R. Pati and N. Fuhrman, "Densification, Swelling and Microstructures of LWR Fuels Through Extended Burnups," Proceedings of ANS Topical Meeting:
LWR Extendri Burnup Fuel Performance and Utilization, April 4-8, 1982, Williamsburg, Virginia.
13 S. J. Dagbjartsson, et al., " Axial Gas Flow in Irradiated PWR Fuel Rods,"
TREE-NUREG-1158, EG&G Idaho, Inc., Idaho Falls, Idaho, September 1977.
l 18' E. Roberts, et al., Section on Point Beach 1 Fuel Rods in " Determination and Microscopic Study of Incipient Defects in Irradiated Power Reactor Fuel Rods," V. Pasupathi and J. S. Perrin, EPRI NP-P12, Electric Power Research Institute, Palo Alto, California, July 1978.
15 J. T. A. Roberts et al., " LWR Fuel Performance Program: Progress in 1978," EPRI NP-1024-SR, Electric Power Research Institute, Palo Alto, California, February 1979.
16 W. J. Leech and R. S. Kaiser, "The Effects of Fission Gas Release on PWR Fuel Rod Design and Performance," Westinghouse Electric Corporation, Pittsburgh, Pennsylvania, Paper Presented at the IAEA Specialists Meeting on Water Reactor Fuel Element Modelling, Blackpool, U.K., March 16-21, 1980.
17 H. Ocken and J. T. A. Roberts, " Comments on Fission Gas Release from Fuel at High Burnup," Letter to the Editor, Nuclear Safety, 20 (July-August 1979),
No. 4, p 417.
18 R. G. Bellamy and J. B. Rich, " Grain-Boundary Gas Release and Swelling in High Burnup Uranium Dioxide," J. Nucl. Mat., y (1969), p 64, 19 C. E. Beyer and C. R. Hann, " Prediction of Fission Gas Release from UO2 Fuels," BNWL-1875, Battelle Pacific Northwest Laboratories, Richland, Washington, November 1974.
20 R. O. Meyer, C. E. Beyer, and J. C. Voglevede, " Fission Gas Release from Fuel at High Burnup," NUREG-0418, U. S. Nuclear Regulatory Commission, Washington, D.C., March 1978.
21 R. Hargreaves and D. A.
Collins, "A Quantitative Model for Fission Gas Release and Swelling in Irradiated Uranium Dioxide," J. Br. Nucl. Energy 1_5_ (1976), p 311.
5 Soc.,
22 T. E. Hollowell, P. Knudsen and H. Mogard, "The International OVER-RAMP Project at Studsvik," Proceedings of ANS Topical Meeting: LWR Extended Burnup Fuel Performance and Utilization, April 4-8, 1982, Williamsburg, Virginia.
e O
r Table 1.
Key Design Parameters, Op,2 rating Characteristics, and Fission Gas Release Results for Test Fuel Rods From Calvert Cliffs I 1
NB0 a
Fuel Rod Numbers 01 05 11 12 09 144 50 58 53 54 60 23 33 46 47 39 42 48 Fuel Parameters i
DENSIFYING
- tt0N-DENSIFYING A
C s0N-DENSIFYlMC M Type I
II II II II III IV IV V
t 1 TD 1
93 e-95 95 95 95 93 95 95 c
93 Wt. 1 U-235 4
2.45 2.45 e 2. 33 +
0 2.82-5 Foreformer 4
NOME t
FVA FVA FVA FVA letA S-K S-K t
FVA 7
Initial Crain Site, um 2.5 4
4 4
4 7
4 4
4 15 t
Open Forootty, 2 of Fellet Volume 0.1 1.3 1.3 1.3 1.3 2.2 0.35 0.35 4
3.7
_Operatina Parametere Number of Cycles 1
2 3
3 4
4 1
2 3
4 3
4 4
1 2
3 3
4 Feak L.HCR, kw/ft. (in Cycle 1) 4 9.1 4
9.8 i
8.8 8.8 4
II.I i
Rod and Time-Averaged IJtGR, kw/f t Cycle I 5.5 5.5 5.5 5.5 5.5 5.6 5.5 5.5 5.5 5.5 5.5 5.3 5.3 6.3 6.3 6.3 6.3 6.4 2
5.1 S.I 5.0 4.9 5.4 5.0 5.1 5.0 5.0 4.9 4.9 5.3 5.3 5.4 5.5 3
4.3 4.4 4.4 4.0 4.3 4.3 4.4 4.4 4.4 4.7 4.6 4.5 4
3.7 4.0
- 3.7 3.8 3.8 3.9 Rod-4veraged Burnup at Discharge, CWd/stu 18.7 25.8 33 33 4I.4 42.2 18.7 25.8 33 41.4 33 40.9 40.9 28.6 29.1 37 37 45.8 b
Fission cas Release, 1 0.27
- 0. 34 0.36 0.35 0.36 1.45 0.33 0.35 0.33 0.55 0.59 0.37 0.28 0.78 0.64 0.78 0.72 0.93
- Common Design Parameters (Nominal)
Fuel Column 1,ength, inches 136.7 Cladding ID, inchee 0.388 Initial Fill Cas Pressure - 450 pois Eacept Rods 23 Fuel Rod I.ength, inches 147 Fuel Rod OD, inchee 0.440 and 39 which had 300 pais @ 20C Fuel Fellet OD, inches 0.3795 Fellets Dished at Both Ende Initial till cas Composition - Nettum Except Rod Assumes a Production Rate of 30 Atome of Ie + hr/100 Fissions and 200 MeV/ Fission
8 s
Table 2.
Condensed Temperature History of Fuel in Selected Calvart Cliffs I Rods Rod Identity 48 54 09 Fuel Type Nondensifying (V)
Nondensifying (II)
Densifying (I)
Enrichment, wt% 235U 2.82 2.45 2.45
-Cycle 1, at Local Peak Heat Rating:
LHGR, kw/ft 11.1 9.2 9.2 2
Gap Cond., w/cm -C 0.8 0.6 0.4 Fuel ( Temp.
C 1490 1270 1400 Fuel Surf. Temp., C 520 530 610
-Cycle 1, at Rod-Avg.
Heat Rating for Cycle:
LHGR, kw/ft 6.3 5.4 5.4 l
2 Gap Cond., w/cm -C 1.5 1.0 0.6 Fuel ( Temp., C 810 730 780
-Cycle 2, at Rod-Avg.
Heat Rating for Cycle:
LHGR, kw/ft 5.5 5.0 4.9 2
Gap. Cond., w/cm -C 3.0 2.0 1.0 Fuel ( Temp., C 700 660 680
-Cycle 3, at Rod-Avg.
Heat Rating for Cycle:
LHCR, kv/ft 4.4 4.3 4,4 2
Gap cond., w/cm -C 3.0 3.0 1.5 Fuel ( Temp., C 610 590 630
-Cycle 4 at Rod-Avg.
Heat Rating for Cycle:
LHGR, kw/ft 3.9 3.7 3.7 2
Gap Cond., w/cm -C 3.0 3.0 2.0 Fuel ( Temp., C 570 540 550
8 s
Figure 1.. Gas Release Measured in Calvert Cliffs I Fuel Rods i
m a
o.
.. a.
O "
e I.
,,O e
..mu.
=
l:
O. Lu.
S-
^
entre
_g_.-.-- ms a
9 8 --
m a
=
m a
u.
t Figure 2.
Gas Release Data From PWR Fuel Rods Compared With the Low-Temperature Data of Bellamy and Rich 8
5 s
a s
a f ut L CthTER flasp. I Q OR R00 fiest AVERACE SVnA00 L SOUACE MEAT RATihC EWrf f 7 = g CALVERT Ctif FS 1 4844 RWST
=
0 "a noeins0mu 6182 seat 18 0AT A P014TS)
D' OC04EI III att 0 AT A P0thill g, G 7014T SE ACM 1(2 CYCtt> I4 8245 EW41 is O
PC'*t et ACn i ta CvCtti o
z0 nit A atoW Poetniil assa swar a
etttAuvs niCu'8
<tles C
O I.
5 a
3 Oo 2
0 O
e== =.
00 o
1 e
00 e O g
3 E Eg 390 0@
0 e
e 18 28 3B 48 88 Bu ANUP, CWd48TU
s 22 Figure 3.
Fission Gas Release From Over-Ramp Rods I
I I
I I
ROD AVERAGEb
- suRauR,
/
swamTu~
O/
O 12.is Spa O A O asas o
h 13 10
]
O su=
l at, 7
29 31 5
0 5
/
2 4.s p a p
.jh 4.5 pm i
le
/
10.5 p m 0
o }"""
V 0
0 en s i
s, 12 13 14 Il 18 17 R AAAP TERailh AL LINE AR HE AT R ATE, NW/FT Figure 4.
Centerline Temperature in Solid Pellet UO2 Fuel as a Function of LHGR for Various Assumptions on Gap Conductance
==
i i
... p...m.a.,
/
MeerYSAtu C a
,, u sus t.
=
orse 3
I i
[
isse i
E 3
l I
ne h
' ' ~ $tn'aI'e's'. art s
l E
I im i
l i.
u i.
i.
is l
...u...,
l I
i s
DENSIFICATION, SWELLING AND MICR0 STRUCTURES OF LWR FUELS THROUGH EXTENDED BURNUPS s
S. R. Pati and N. Fdirman Combustion Engineering, Inc.
- Nuclear Power Systems Windsor, Connecticut l
Under the joint sponsorship of the Electric Power Research Institute (EPRI) and Combustion Engineering (C-E) the densification and swelling behavior of UO Pellet fuel, exposed for up to 4 cycles in Calvert Cliffs-I, 2
has been characterized and correlated with pellet microstructure and fuel i
rod performance. Data are included from threa different fuel types, two nondensifying (93% TD and 95% TD) and one densifying (93% TD) operated under typical PWR conditions, i.e., centerline temperatures of less than 1300C Irradiated fuel specimens, representing local burnups estimated y37 from Cs gamma scans, and ranging to 50 GWd/mtU, weta subjected to mercury pycnometric measurements. The resulting density data were related to the respective as-fabricated densities to generate sets of data on volume change vs. burnup for each fuel type. Because of early-in-life densifica-tion, a clear indication of the swelling trends was evident only above a burnup of 20 GWd/mtU. Linear regression analyses on the data sets above this burnup were performed to compare the indicated swelling rates with matrix (i.e., N1.0% volume change that published for the pore-free UO2 per 10 GWd/mtU). These evaluations demonstrate that densification affected the net swelling behavior of the densifying fuel up to a burnup of approxi-mately 35 GWd/utU, above which the fuel exhibited the matrix swelling rate.
The trends for the nondensifying fuels indicate, that at high burnup, swelling was accommodated within large open pores in the microstructure l
such that the external swelling rate was less than the, matrix swelling rate. A comparison with the corresponding fuel rod performance data i
indicates that this behavior was related to Zircalo'y cladding restraint.
Similar regression analyses on data obtained du' ring an EPRI-sponsored l
post-irradiation examination of two high burnup Connecticut Yankee fuels, also operated at less than 13000, show agreement with the matrix swelling rate indicating an absence of the internal swelling accommodation that was observed in the case of the nondensifying Calvert Cliffs-I fuels.
O 4
i
' s DENSIFICATION, SWELLING AND MICR0 STRUCTURES OF LWR FUELS THROUGH EXTENDED BURNUPS INTRODUCTION In the late 1960's, a gradual shift took place toward the use of a relatively low density UO, as a commercial light water reactor (LWR) fuel, because of the long-standing belief that porosity distributed within the UO matrix is effective in accommodating irradiation-induced swelling of fuel.2 l
The lowering of fuel density, without appreciation for the role of pore size distribution, led to the early 1970's problem of cladding collapse into fuel In re8ponse to the l
column gaps caused by in-reactor densification of UO2 fuel densification problem, manufacturing modifications were introduced by l
fuel suppliers to obtain a DO, pellet microstructure with a large fraction of 3
the pore volume in the form of large pores. As a result, the new fuels showed minimal densification under irradiation '". Subsequently a coordinated 3
program on fuel densification, sponsored by the Edison Electric Institute (EEI) and the Electric Power Research Institute (EPRI), was undertaken by the LWR industry to obtain a better understanding of the phenomenon of fuel densification. As the newer fuels were introduced, the specified fuel s
densities were also generally raised to the level of 95% of theoretical.
These developments were accompanied by a greater awareness that the not fuel volume change under irradiation represents the summation of a low-burnup densification cogonent and a superimposed swelling component which dominates at higher burnups '7 The interplay between these two processes, interacting with. cladding creepdown, was recognized as a key factor underlying the fuel 6
rod thereci and mechanical behavior throughout life. For example, some swelling accommodation is afforded late in the life of a fuel rod, depending upon the degree of densification, early-in-life. It is also recognized that a part of the late-in-life swelling could be accommodated by the pores within the fuel. Thus, with the advent of ef forts to improve uranium utilization by extending the discharge burnup of UO, fuel, there is heightened interest in l
how swelling is accommodated in modeh fuels, especially at high burnup.
Swelling measurements on a cumber of different fuel types had been reported in the past by Daniel, et al.8 and others.9,10 However, some 1
c of these investigations,10 were performed before densification was widely 8
recognized, and did not clearly account for the effects of densification on In other work '8, the fuel swelling was inferred l
li the overall volume changes.
from measurements of fuel column length changes, which may not provide a good
'\\
measure of the net volumetric swelling rate because of the complications introduced by (1) pellet cracking and radial relocation and (2) fuel creep which.makes swelling anisotropic under restraint from the cladding.
,,.-,._c
-.,n
.n.
,,e-
-,,--.-_n.
i With a better understanding of the role of densification, Assmann and Manze17 used mercury pycnometry to measure the net swelling rate of a commercial UO2 pellet fuel irradiated under thermal conditions typical of LWR operation, i.e.,
<1300C centerline temperature. Under these conditions, swelling is induced only by the generation of solid fission products plus gaseous fission products retained within the UO, grains in small bubbles or in solution. In the burnup range of 15 to 45 GWd/mtU, a not swelling rate of 110.15% volume change per 10 GWd/mtU was obtained. Using quantitative I
ceramography, the swelling rate of the pore-free UO2 ***#A* "** I ""d A" agreement with the net rate derived from the data using mercury innersion techniques. These results indicated that, for the type of fuel studied, (1)
., the annihilation of small pores by the densification process was essentially complete below a burnup of 15 GWd/mtU, and (2) the remaining as-fabricated pores were sufficiently large to be stable under further irradiation. In reviewing the volumetric swelling rates of UO btained by previous workers 2 7 for this temperature regime, Assmann and Manze1 noted that all rates were equal to, or lower than, the matrix swelling rate determined in their l
investigation. As discussed in this paper, the pore-free matrix swelling
[
rate is the upper limit, as both the densification and internal accommodation t
reduce the rate to lower values. These values can vary widely, depending on the particular microstructure of the UO, the extent of cladding restraint
(
3 l
and the operating conditions of the fueI rod.
The Fuel Performance Evaluation Program in Calvert Cliffs-I (CC-I),
f jointly spor.sored by EPRI and Combustion Engineering (C-E), has provided an opportunity to obtain densification and swelling data on three fuel types, irradiated under typical operating conditions of power reactors. The major objective of the overall program is to evaluate the performance of test fuel rods which have systematic design variations in properties such as fuel microctructure, initial fuel density and propensity for in-reactor densification.
The fuels were fabricated in the " post-densification" period and differences in in-reactor densification propensity wetre specifically included in the test 4
matrix so that the effect of this parameter could be evaluated under power reactor conditions. Data obtained from this program include fuel rod dimensional changesll~14, void volume and microstructural evolutionlb~as well 17 as fission gas release after two, three and four cycles in-reactor I
Densities of the irradiated fuel pellets have now been measured by mercury pycnometry at various burnups up to %50 GWd/mtU. These data allow a comparison of the densification and swelling behavior of the three micro-structures as a function of burnup.
The results of the EPRI/C-E program are evaluated in this paper, together with the results of similar density measurements on two other fuel types irradiated to approximately 40 CWd/mtU and examined as part of an investigation of fuel rod performance in the Connecticut Yankee (CT) Reactor 18, That program was sponsored by EPRI and Northeast Utilities Service Company (NUSCO).. The fuels examined under that program were irradiated in Batch 7 and Batch 8.
The Batch 7 UO, was a densifying fuel type, though its as-was a nondensifying fabricated density was 95% TD; while the Batch 8 UO2 fuel type having the same starting density.
l t
e 1
EXPERIMENTAL PROCEDURES The UO Pellets examined in this investigation were irradiated in test 2
I rods which were part of a surveillance program at CC-I.
The history of these fuel specimens, the basis for their selection, and the procedures followed for density measurements and calculation of volume changes are briefly described below.
SPECIMEN HISTORY Post-irradiation densities of three fuel types were measured as a function of burnup ranging from 9 to 50 GWd/stU. The fuel types investigated included a densifying fuel and two types of nondensifying fuels. Variations in initial density, porosity distribution and grain size were present in the fuel types examined. The pettinent fabrication parameters and attributes are summarized in Table 1.,As-fabricated porosity structures of the fuel types are reproduced in Figures la through ic.
UO specimens representing various burnups were obtained by taking y
samples from fuel rods that were irradiated in CC-I for two, three and four cycles of exposure. The test fuel rods used are identified in Table 1 together with the important operating parameters. The rods operated under normal duty cycles typical of power reactor operation. The cycle-averaged heat ratings of the rods ranged from 3.7 to 6.4 kw/ft and the heat rating generally decreased with time. Early in the first cycle, the average heat ratings differed by %1 kw/f t due to differences in enrichment. However, as irradiation progressed, heat ratings of all of the rods approached each other and were essentially identical towards the end-of-life. The operating history of these rods are described in greater detail by Bessette, et al.ll and Ruzauskas, et al.12 13, SPECIMEN SELECTION Initially, specimens at three different burnups were obtained from a single test rod by selecting duplicate fuel pellet samples from euch of three different axial elevations of the same rod representing nominally high, intermediate and low levels of burnup. Additional samples were later selected from some of the three-and four-cycle rods containing nondensifying fuels to verify and improve the data correlations. Sample locatio g for density measurements were selected ef ter reviewing the available Cs gamma scans of the test rods 15,16 The axial locations of the selected pellet specimens were eosily distinguished due to the sharp reduction ofggtivity at the pellet interfaces. To estimate the local burnup, the average Cs activity was first determined for each rod by an integration procedure applied to the respective ganna scan. Using the Pyysics calculated rod average burnup for y
each rod, the ratio of burnup to Cs activity was then established. Local burnups we g calculated for each pellet specimen by applying this factor to the local Cs activity determined from the respective gamma scan. Burnups calculated by using the above procedure for fuel samples selected from Rods 11, *2, 47, 51 andy were found to sgree (within 4%) with the burnup determined chemically by the Nd method 15 16,
t I
DENSITY MEASUREMENTS BY THE MERCURY PYCNOMETRIC TECHNIQUE Densities of the irradiated UO samples were determined by the mercury 2
pycnometric method. The specific procedure used was identical to that used by Brite, et al.5 and, in both cases, the measurements were made at the Battelle-Columbus Laboratories. The procedure used is applicable to measuring the bulk density of regularly or irregularly-shaped fuel material. In all cases, the samples were representative of an entire fuel pellet cross section.
l with a specimen length equivalent to a pellet length. Sampling of the irra-l m
diated fuel was accomplished by removing the fuel from the cladding, generally by carefully pushing it out of a transverse section. For some of the samples.
it was necessary to slit the cladding with an abrasive cutting wheel and pry L
the fuel loose. The samples consisted of all the fuel from a designated section of the fuel rod, minus fuel particles of less than 10 mesh (1.68 mm diameter). The entire technique was checked daily by measuring the density of a stainless steel standard. The accuracy of the technique was established to be within +0.02 g/cm3 from the true bulk density.
CALCULATION OF VOLUME CHANGES INDUCED BY IRRADIATION For a comparison with published volumetric swelling data, e.g., the l
UO matrix swelling rate, the net volumetric changes (% AV/Vo) resulting from 2
the irradiation were calculated from the densities associated with the as-fabricated and the irradiated states.
In the case of the nondensifying fuels, individual as-fabricated 19 The geometric densities were available for 24 of the pellet samples initial geometric densities for the remaining nondensifying pellet samples were assumed to be the mean density determined for the respective fuel type during the precharacterization program.
[The standard deviations are in the range of 0.3 to 0.4% TD (0.03 to 0.04 g/cm ) for both fuel types.]
3 densities In the case of the densifying fuel, the initial average UO2 of the respective fuel rods from which the samples were taken were used as the as-fabricated geometric densities of the samples. These were computed from the known fuel stack weights and dimensions, using the known stack lengths, nominal pellet diameters, and dish volumes.
For computing the irradiation-induced volume change, the as-fabricated geometric densities were corrected for the inherent bias between the densities determined by the geometric and mercury pycnometric methods. The applicable corrections for each of the fuel types were determined by measuring the densities of whole pellets in the as-fabricated s. tate by both methods.
These measurements were available on two archive fuel pellets each of the Fuel Types I and V and on one fuel pellet of the Fuel Type II. For the Fuel Types I, II and V, the mercury imunersion densities were higher than the 3
respective geometric densities by 0.056, 0.063 and 0.133 g/cm, respectively.
a RESULTS AND DISCUSSION VOLUME CHANGES OF CALVERT CLIFFS-I FUELS The irradiation-induced volume changes, calculated for the three fuel types from CC-I, are plotted as a function of burnup in Figure 2.
The data show that, at the lowest burnup examined (%10 Wd/mtU), densification was dominant in the densifying fuel and in one of the nondensifying fuels. The densifying fuel, as expected, showed a much larger volume decrease than the nondensifying fuels. Of the two nondensifying fuels, the fuel with the lower initial density (93% TD) showed a net volume increase in this low burnup
- region, compared to a net volume decrease exhibited by the higher density (95%
TD) fuel. A net volume increase at such a low burnup is indicative of a relatively small amount of densification to compensate for the fuel swelling.
Some caution needs to be exercised, however, in interpreting these low burnup data. Samples taken from the lowest burnup regions operated at lower tempera-tures than those from the high burnup regions of a given rod. Such low burnup samples may have experienced a reduced densification rate. Temperatures in this range have an effect on densification rates, and the temperature of these samples was below the threshold for temperature-independent densification, N700-750C.20 Lower fission rates may also have contributed to a lower 5
extent of densification in these low burnup specimens. Only above this threshold temperature may densification be considered athermal, independent of fission rate 20 and dependent only on total fissions (burnup).
Although the interpretation of the low burnup data must be tempered by this temperature and/or fission rate effect on densification cates, the data generally show that the densification component exerts a major influence on the overall volume change of these fuels to burnups of at least 20 Wd/mtU.
The contribution from the densification component delayed the development of a clear trend of swelling-induced volume expansion with increasing burnup.
Only at burnups above 20 Wd/mtU, does the swelling component become dominant and leads to continuous volume increases with increasing burnup. A cursory examination of the data in Figure 2 in this higher burnup region indicates that two distinctive volume change trends are being exhibited by these fuels.
The indicated rate of volume increase for the densifying fuel appears to be increasing with burnup. In contrast, the volume change data for the non-densifying fuels indicate that a transition to a decreased rate of volume increase may be occurring at high burnup. These qualitative trends are apparent in Figure 2.
However, there are finer differences within these trends, which are the subject of detailed discussion in the next section.
If, indeed, the small pores have been annihilated by in-reactor densi-fication and the remaining pore size distribution is unchanged above 20 Wd/mtU, l
the corresponding volume changes should result in the swelling rate reported 7
m8trix, Perating below temperatures by Assmann and Manze1 for a pore-free UO2 associated with fission gas bubble swelling. The thermal history of the fuels from CC-Il7 indicates that, except for a brief period early in the first cycle, the fuel centerline temperatures ranged from 600 to 800C. This range is well below the threshold temperature for fission gas bubble swelling. Ceramographic examinat. ions of a number of fuel cross sections also revealed no appreciable 15 16 Thus, porosity a.ccumulation due to coalescence of gaseous fission products the swelling must be entirely due to solid and gaseous fission products retained within the fuel and dispersed on a submicroscopic scale.
.,---._.v.
n--,-
_.3
l s
s i
DATA ANALYSES With the assurance that the thermal histories were comparable, a series of linear regression analyses-were performed on the data in Figure 2.
Burnups above 20 GWd/mtU were used for each fuel type to minimize the e,ffects of densification. These analyses yielded rates of volume increase for com-parison with the Assmann and Manzel swelling rate for the matrix.
It needs to be emphasized, however, that volume changes derived frca mercury immersion techniques reflect only the changes in the external volume of the fuel, as s
the mercury does not penetrate the porosity within the fuel. Therefore, the density measurement does not account for any additional swelling into the porosity within the fuel (internal swelling). In contrast, the Assmann and Manzel pore-free matrix swelling rate of UO2 provides an upper limit on swelling, as they specifically treated the data using quantitative ceramo-graphic information to exclude the contribution from internal swelling. Also, their experimental observations did not reveal any appreciable chrinkage of the coarse porosity that could have been caused by internal swelling or densification.
It is expected, therefore, that the net swelling rates deduced from the mercury immersion data could be lower than the pore-free matrix swelling rate (even in the absence of densification) and would depend somewhat on fuel microstructure and the degree of external restraint on the fuel.
The key results of the regression analyses are presented in Table 2 and the resulting linear relationships are shown in Figure 2.
The indicated densification and swelling behavior is highlighted below for each of the' fuel types.
A.
Densifying Fuel (Type I)
In the case of the densifying fuel of 93% TD, there is a high correlation coefficient (0.93) indicating a good linear fit between the volume change and burnup data.
The indicated swelling rate is 0.75% volume change per 10 GWd/mtU, which is somewhat lower than the rate reported by Assmann and Manzel. Since this was a densifying fuel, one obvious explanation is that the volume changes at higher burnups still included a significant contribution due to densification (i.e., annihilation of remaining pores was continuing at a measurable rate),.
As a check on the above possibility, another regression analysis was performed, this time restricted to data above 35 GWd/mtU. The resulting indicated swelling rate increased to 0.99% volume change per 10 GWd/mtU, in better agreement with the pore-free matrix swelling rate of UO. An examinationoftheas-fabricatedporositystructureofthisfuelfFigure
~
1a) revealed a range of sizes of fine pores. It is apparent that some of the larger pores still within the densifiable range were present at burnups in the 20-35 GWd/mtU range and provided a densification component 20 to the volume changes.
It is well recognized in the literature that the densification process may proceed over an extended burnup range, depending upon the pore size distribution present in the fuel.
The indicated swelling rate above 35 GWd/mtU almost reached the value for the swelling rate associated with the pore-free matrix. This is evidence that no internal swelling accommodation is occurring at high burnup in the densifying fuel. Such swelling accommodation is not considered likely in this fuel type for the following reasons:
(1) The fuel reached a reasonahly high density early-in-life, so that only a small amount of porosity remained at the time when swelling became dominant; (2) essentially all of the as-fabricated pores were small and closed, so that they were likely to have developed high internal pressures due to fission gas accumulation, and as a result, were resistant to swelling forces; and (3) the fuel operated with large gaps and did not experience significant restraint from the cladding, as is evident from
~
the measurements of, fuel rod ovality and diameter obtained by profilometry at the spent fuel pooll l l. These gaps were also observable on metallographic samples 15,16, B.
95% TD Nondensifying Fuel (Type II) f The analysis of the 95% TD nondensifying fuel yielded a high swelling rate of 0.92% volume change per 10 GWd/mtU with a correlation coefficient of 0.94.
This swelling rate is in good agreement with the pore-free matrix swelling rate, so that a concurrent densification process does not appear to contribute ::f snificantly to the overall swelling behavior of this fuel type at burnups above 20 GWd/mtU. This behavior is consistent g
with the fuel microstructure shown in Figure Ib and the therinal resintering test results included in Table 1.
The fuel contained a small amount of densifiable porosity which was essentially annihilated within the first 20 GWd/mtU.
To obtain an indication of the degree of internal swelling accommodation, if any, that may have concurrently operated in this fuel, a linear regression analysis was performed on a subset of the data above 35 GWd/stU.
As shown in Table 2, this yielded a swelling rate of 0.80% volume change per 10 GWd/mtU (with a correlation coefficient of 0.94). This result suggests that internal swelling acconmodation equivalent to 0.2% volume change per 10 GWd/mtU may have operated in this fuel type at burnups above 35 GWd/mtU. The dimensional data from the fuel rods using this fuel type, as well as the observed fuel-cladding gaps on taetallographic samples 16, suggest that some cladding restraint developed on this fuel at these higher burnups. Based on measurements of fuel rod void volume on this rod, 45% of the total porosity of the pellets was in open pores 16 These interagglomerate, concentrated on the' agglomerate boundaries open pores appear to promote internal swelling accommodation under adequate 4
restraint from the cladding at high burnups. The capability of the coarse, interagglomerate open pores to accommodate swelling is discussed below in more detail, while evaluating the behavior of the lower density, nondensifying fuel.
8 s
C.
93% TD Nondensifying Fuel Crype V)
The analysis of the data from the 93% TD, nondensifying fuel yielded a much lower indicated swelling rate (0.4 to 0.5% volume change per 10 GWd/mtU). Although the three-cyc7.e rod (#42) contained the same fuel type as used in the two-cycle (f47) and four-cycle (#48) rods, the UO2 pellets were from different lots. The fuel lot used in Test Rod #42 appeared to have densified more than the fuel lot used in Test Rods #47 and #48. As a result, as the irradiation continued, the volume of the fuel from Rod #42 remained somewhat lower than that of the fuel used in the other test rods. Consequently, a separate regression analysis was performed on the data generated from Test Rod #42. The results show a relatively Iow swelling rate of 0.43% volume change per 10 GWd/mtU with a reasonable linear correlation.
To improve the correlation, as well as to verify the low value of indicated swelling rate exhibited by the other lot of this fuel type over an extended burnup range, additional density data were obtained from the two-cycle and four-cycle test rods overlapping the burnup range of the three-cycle test rod. The regression analysis of the data from these test rods yielded a swelling rate of 0.48% volume change per 10 GWd/mtU, with a correlation coefficient of 0.87.
Two possible mechanisms need to be' considered to account for this behavior:
(1) continued densification at burnups above 20 GWd/mtU and (2) internal accommodation of matrix swelling within the pellets. Both of these lead to a lewer ntt indicated swelling rate compared to the swelling rate of the pore-free matrix. An indication of the relative densification propensity of this fuel, obtained from the thermal resintering tests, showed that this fuel is likely to exhibit a minimal amount of densification under irradiation. Unlike the densifying fuel, essentially all of the porosity in this fuel was very coarse and was concentrated in the interagglomerate regions (Figure ic). This fuel had a high level of open porosity (70% of the total porosity) due to the interconnection of the interagglomerate regions with the surface.
In contrast, nearly all of the total porosity in the densifying fuel was closed.
The irradiated structure of the Type V fuel used in the two-and four-cycle rods is presented in Figure ld for comparison with the as-fabricated structure shown in Figure Ic. A significant reduction of the coarse porosity due to irradiation is evident. Although it is not possible to rule out a reduction in porosity due to continued densification, an examination of the general trend of the reduction of the fuel rod internal l
void volume with burnup strongly suggests that densification is not a major factor. Specifically, it is recognized that changes in internal void volume of fuel rods with irradiation result from a superposition of net external volume changes in the fuel (combination of densification and swelling), cladding dimensional changes, filling of open porosity due to ovelling accommodation and any additional change in open porosity content caused by processes other than swelling accommodation. Changes in fuel rod void volume resulting from the dimensional changes of the cladding could be estimated from dimensional data obtained at the spent fuel I
l
i poo111~13 With the assumption that the additional changes in open porosity (not caused by swelling) were negligible, changes in void volume attributable to external volume changes of the fuel and internal swelling into open pores could be estimated. Such computations indicated that the internal void volume of the roda van being reduced at a rate I
which corresponds to about 1% volume change of the fuel per 10 GWd/mtU.
Thus, the total swelling rate (external volume change plus internal swelling into open pores) agrees reasonably with the swelling rate of the pore-free matrix. This analysu supports the hypothesis that densification is not contributing to the observed swelling behavior of this fuel at high burnups.
In summary, the above data evaluation indicates that the low net swelling rate of the 93% TD, nondensifying fuel (and perhaps that of the 95% TD non-i densifying fuel as well) resulted primarily from the accommodation of part of the pore-free matrix swelling into the open pores, which were preferentially distributed in the interagglomerate regions. The profilometry measurements on the associated fuel rods, as well as the fuel-cladding gaps observed on tne metallographic samples are qualitatively censistent with the above explanation.
2 Only those fuel rods using this fuel type showed an increase in the fuel rod diameter with increasing burnup starting from the end of the second irradiation cycle (rod average burnup of 29 GWd/stU). Also, the fuel-cladding gaps were 15 at the end of the second cycle. Thus, the fuel operated essentially closed under cladding restraint, which provided the necessary driving force for part of the swelling to be directed to the internal open pores.
It is of interest that a mixed-oxide fuel with a structure similar to that of Fuel Type V has been developed in the FFTF program 21, using a high pressure preslugging technique. This techntque produced fuels with a low quantity of fine porosity (<2 um) that was attributed to the very hard dense granules that sinter well. Large intergranular pores resulted because of loose bonding between the granules. The open porosity content of this fuel is Hanford Engineering Development Laboratory (HEDL)21, previous experience, a also high, and similar to that of Fuel Type V.-
The as well as the results of other densification studies 3 5, suggest that a significant removal of such coarse open porosity by pore annihilation is unlikely. This HEDL fuel develop-ment effort, and the performance of the Type V fuel in a PWR, suggest a future direction for advanced LWR fuel development intended for high burnup applications.
In addition to the various net swelling rates obtained for the different
~
fuel types at high burnups, the regression analyses yielded intercepts on the AV/V, axis from the extrapolation of the linear relctionship of AV/V, with burnup. These intercepts, which are included in Table 2, reflect the relative densification behavior of the different fuel types. For the CC-I fuels, the magnitude of the intercepts indicate a relative densification tendency of f
93% TD densifyingn95% TD nondensifying>93% TD nondensifying.
This is in agreement with the expected ranking from the previously reported resintering test results as listed in Table 1 and the porosity structures of the fuel types shown in Figures la through ic.
,y
_ _ ~. _ _ _. -
ANALYSIS OF CONNECTICUT YANKEE FUEL DATA Mercury pycnometric density measurements were similarly conducted as a function of burnup on samples from two hatches of fuel from CY as part of the EPRI/NUSCO-sponsored fuel performance investigationl8 The Batch 7 UO2 "*
- a densifying fuel type, while the Batch 8 UO, was a nondensifying fuel type.
Both fuels had as-fabricated geometric densiEies of 95% TD and the burnups achieved were up to 44 GWd/mtU. The fuel rod design for each batch was the same, and the cladding for these rods was stainless steel, rather than Zircaloy-4 as in the CC-I rods.. The peak burnup reached in Batch 7 was slightly higher than in Batch 8; but, in both cases, the linear heat ratings
- were typical of power reactor irradiations, so that the centerline tempera-tures of the UO were generally less than 1300C. From the reported data, 2
volume changes were determined as a function of burnup in the range above 20 GWd/mtU. Linear regression analyses were performed and the results are 1
shown in Table 2.
It should be noted that data for a correction of the bias between immersion and geometric densities were not available. This correction, however, would only affect the intercept values at zero burnup.
The high correlation coefficients obtained (>0.97) indicate a good linear fit between the volume change and burnup data for both fuel types.
More importantly, both fuels exhibited swelling rates in reasonable agreement with the UO matrix swelling rate determined by Assmann and Manzel so that 2
densification does not appear to be contributing significantly to the overall swelling behavior of these fuels at high burnup. However, the difference between the intercepts in Table 2 (even with the uncertainty of the bias) are indicative of a substantially greater early-in-life densification in the Batch 7 fuel than in the Batch 8 fuel. Thus, as burnup continued, the bulk density of the Batch 7 fuel was always at least 1% greater than that of the Batch 8 fuel. Evidence that such densification was beneficial in providing seelling accommodation in the CY fuel rods was obtained from other post-irradiation examinations of Batch 7 and Batch 8 fuell8 Profilometry of the highest burnup rods indicated less cladding ovality and more ridging in Batch 8 than in Batch 7.
Also, the fuel-clad gap was closed in the Batch 8 metallographic samples, while a small gap remained in a Batch 7 sample.
It should be noted that the above observations concerning the swelling accommodation characteristics of the Batch 8 UO fuel are consistent with the 2
22 behavior of other fuels with a similar microstructure. Raven has reported that the large closed pores (typically 20-40 pm) in such fuels are effective in accommodating swelling by closing under cladding restraint at fuel tempera-tures above 1300C; however, below this temperature, these pores remain unchanged.
Although the fuel was restrained by the cladding in the Batch 8 rods near end-of life, the fuel temperature apparently was too low to obtain significant swelling acconunodation in the large closed pores.
g o J i
4 CONCLUSIONS The resulta of the fuel density measurements from CC-I and CT fuels-lead to the following conclusions:
The three fuel types investigated from CC-I exhilind di9tinctly different densification and swelling patterns. In the densifying fuel, pore annihilation continued up to 35 GWd/mtU, though at a rate much slower than at the beginning-of-life. The concurrent densification process was responsible for the swelling rate remaining below the 1% volume change per 10 GWd/mtU that has been reported for the pore-free UO matrix.
In the nondensifying fuels, 2
densification was essentially completed by a burnup of 20 GWd/mtU.
With the nondensifying fuels from CC-I, operated under the duty cycles typical of normal operation in a PWR, there is evidence that swelling accommodation occurs within the open pores of the UO2 i
microstructure, and that the amount of accommodation depends upon t
the restraint of the Zircaloy cladding. Swelling accommodation into I
the pores keeps the observed swelling rates below the rate for the f
pore-free matrix, even in the absence of densification.
In the densifying fuel, the observed swelling rate after 35 GWd/stU was in excellent agreement with the matrix avelling rate for pore-free UO, because of the absence of swelling accommodation up to g
44 GWd/mcU.
L The 93% TD nondensifying fuel, having 70% of its total porosity i
concentrated on the agglomerate boundaries as open pores, showed the largest extent of swelling accommodation among the fuel types investigated. The observed swelling rate was only half the rate associated with the pore-free matrix. This suggests that tailoring of a fuel microstructure during fabrication could be a viable approach to meeting design objectives which depend on the degree of fuel-cladding interaction desired.
The densification and swelling behavior of the various fuels evaluated, coupled with the respective fuel rod performance data, demonstrates that UO densification and internal swelling accommodation characteris-2 tics are important factors to consider in the designing of fuel rods for extended burnup application.
A comparison of the swelling accommodation characteristics indicates that, at the fuel temperatures corresponding to normal PWR operation, the large open pores of the CC-I nondensifying test fuels have accommodated more swelling than the large closed pores of the CY Batch 8 fuel.
The swelling mechanisms which are operating in the UO at burnup 2
levels to 50 GWd/mtU appear to be gradual, and show no abrupt phenomena which would limit the life of UO fuel rods with Zircaloy cladding.
2 1
.)
i ACICiOWLEDGEMENTS The collection and evaluation of data from Calvert Cliffs-I fuel rods was performed under the joint sponsorship of the Electric Power Research Institute 00PRI Research Project 586-1 Task A) and Combustion Engineering, Inc.
The assistance of Baltimore Gas and Electric's Calvert Cliffs personnel and Battelle-Columbus Laboratories Hot Cell Facility personnel on this task is gratefully acknowledged. Our apprec 1ation is extended to P. Van Saun and other colleagues at Combustion Engineering, Inc. for their contributions to this investigation.
REFERENCES 1
E. Roberts, J. Iorii and R. Argall, " Westinghouse Pressurized Water Reactor Fuel Development and Performance", Nuclear Energy, g,
No. 5, (1980), p 335.
+
2 USAEC Regulatory Staff, " Technical Report On Densification of Light t
Water Reactor Fuels", Washington, D.
C., November 1972.
3 W. R. Yario, et al., "In-Pile Densification Characteristics of Uranium Dioxide", Trans. Am. Nucl. Soc., IJ[ (1974), p 123.
M. G. Andrews, et al., " Evaluating the Solutions to Fuel Densification",
Trans. Am. Nucl. Soc., lj[ (1974), p 140.
5 D. W. Brite, et al., "EEI/EPRI Fuel Densification Project", Final Report, EPRI-131 Palo Alto, California, March 1975.
6 H. Stehle, H. Assmann and F. Wunderlich, UO PropeEties for LWR Fuel 2
Rods", Nucl. Eng. Des., 3jl (1975), p 230 7
H. Assmann and R. Manzel, "The Matrix Swelling Rate of UO ", J. Nucl.
2 Nat., 6_8_ (1977) p 360.
8 R. C. Daniel, et al., " Effects of High Burnup On Zircaloy Bulk UO 2
Plate Fuel Element Samples", WAPD-263, Westinghouse Electric Corporation, l
Pittsburgh, Pennsylvania, September 1962.
9 R. Hargreaves and D. A. Collins, "A Quantitative Model For Fission Gaa Release and Swelling In Irradiated Uranium Dioxide", J. Br. Nucl. Energy Soc.,
15 (1976), p 311.
10 D. Brucklacher and W. Dienst, " Creep Behavior of Ceramic Nuclear Fuels Under Neutron Irradiation", J. Nucl. Mat., 4j[ (1972), p 285.
l 11 D. E. Bessette, et al., " Examination of Calvert Cliffs-I Test Fuel Assemblies at End of Cycles 1 and 2", NPSD-72, Combustion Engineering,
[
Inc., Windsor, Connecticut, September 1978.
i l
12 E. J. Ruzauskas, J. G. Schneider and P. A. VanSaun, " Examination of Calvert Cliffs-I Test Fuel Assembly After Cycle 3", NPSD-87, Combustion l
Engineering, Inc., Windsor, Connecticut, September 1979.
i
S 4 13 E. J. Ruzauskas, J. C. LaVake and R. G. Weber, " Examination of Calvert i
Cliffs-I Test Fuel Assemblies Af ter Cycle 4", C-E NPSD-146, Combustion Engineering, Inc., Windsor, Connecticut, October 1981.
14 M. G. Andrews, S. C. Hatfield and E. J. Ruzauskas, " Operating Experience With Combustion Engineering Fuel at High Burnups", Proc. of ANS Topical _
Meeting on LWR Extended Burnup-Fuel Performance and Utilization, Williamsburg, Virginia, April 4-6, 1982.
. 15 S. R. Pati, "Cas Release and Microstructural Evaluation of One-and Two-Cycle Fuel Rods From Calvert Cliffs-I",' NPSD-75 Combustion Engineering, Inc., Windsor, Connecticut, March 1979.
S. R. Pati, " Gas Release and Microstructural Evaluation of Three-Cycle 16 f
Fuel Rods From Calvert Cliffs-I", C-E NPSD-119, Combustion Engineering,
. Inc., Windsor, Connecticut, December 1980.
17 S. R. Pati, H. R. Freeburn and L. V. Corsetti, " Fission Gas Release From PWR Fuel Rods Under Conditions of Normal Operation and Power Ramping",
Proc. of ANS Topical Meeting on LWR Extended Burnup-Fuel Performance and Utilization, Williamsburg, Virginia, April 4-8, 1982.
V. Pasupathi and R. W. Klingensmith, " Investigation of Stainless Steel 18 Clad Fuel Rod Failures and Fuel Performance in the Connecticut Yankee Reactor," EPRI NP-2119, Electric Power Research Institute, Palo Alto, California, November 1981, 19 V. Pasupathi, " Fabrication and Characterization of BC&E-I Test Fuel Assemblies", Combustion Engineering, Inc., Windsor, Connecticut, f
November 1975.
Densification 20 H. Stehle and H. Assmann, "The Dependence of In-Reactor UO2 On Temperature and Microstructure", J. Nucl. Mat., 5_2 (1974), p 303.
21 D. E. Rasmussen and P. S. Schaus, " Pre-Irradiation Microstructure Characterization of FFTF Mixed-Oxide Fuel", Trans. Am. Nucl. Soc., H (1981), p 372.
22 L. F. A. Raven, "The Performance of BNFL Controlled Porosity Nuclear Fuel", A Paper Presented to the American Ceramic Society 78th Annual Meeting, Cincinnati, Ohic, May 1976, Paper No. 45-N-76 [BNFL Memorandum 469(S), (1976)].
u t
Table 1.
Fuel Types and Test Fuel Rods Examined in the Densification and Swelling Characterization Program Fuel Parameters Type I
II V
Densification Characteristics Densifying Nondensifying Nondensifying-
% TD 93 95 93 235 Wt.%
U 2.45 2.45 2.82 u
Poreformer No Yes Yeg Initial Grain Size,pm 2.5 4
15 Open Porosity, % Pellet 3.7*
i Volume 0.1 1.3 Density Increase During Resintering 0. 0',
Test, % TD 3.0 0.7 Operating Parameters Test Rod No.
05 11 09 51 53 54 47 42 48 No. of Cycles Exposure 2
3 4
2 3
4 2
3 4
Rod Av.Burnup,0wd/MTU 25.8 33.0 41,4 25.8 33.0 41.4 29.1 37.0 45.8 For the lot present in the two-and four-cycle rods.
Table 2.
Results of Linear Regression Analyses 3 Inter-Analyzed cept
- Indicated d
Fuel Parameters Burnup
- Number Corre-on %
Swelling Rate Nominal Densification Range of Data lation AV/V
% Vol. Change o
Type
% TD Characteristic Gwd/MTU Points Coef.
Axis per 10 GWd/mtU I
93 Densifying 23 - 44 12 0.93
-5.36 0.75 I
93 Densifying 36 - 44 8
0.82
-6.35 0.99 II 95 Nondensifying 25 - 45 14 0.94
-2.36 0.92 II 95 Nondensifying 36 - 45 8
0.94
-1.86 0.80 Y
93 N ndensifying 24 - 50 15 0.87
-0.20 0.48 b
V 93 Nondensifying 27 - 41 6
0.76
-0.61 0.43 CY-Batch 7 95 Densifying 27 - 44 10 0.97
-4.09 1.08 CY-Batch 8 95 Nondensifying 22 - 42 8
0.98
-1.97 0.89 a
Lot present in the two-and four-cycle rods.
b Lot present in the three-cycle rod.
Intercepts resulting from linear regression analyses'on the % AV/V vs.
o burnup relationships in Figure 2.
From slopes of the above linear relaticeships.
m m
t-'
r o
Figure 1.
Pore Structures of Fuel Types Irradiated in Calvert Cliffs-I (a) 93% TD Densifying (I) As-Fabricated (b) 95% TD Nondensifying (II) As-Fabricated Longitudinal Section Longitudinal Section I(h:.; ', [.., -..,. ' ' *. -
.. M i ~...'/,.,$,.',,. g
-[,
').. f'.hI ;.' [
.. '.~ '.2 %. +.*.
- ** s
. i e., ?. '.
p. '. ' if. '
.~~
hi?.5.'t.l.'U.d.3,7.';f.s.s." (. M*,.I VMr'.S.,Tz'N6}.h.:.,d.M.'..tO g'.h U h.N, " ' f; '.),
~
'~
.ys.
.!.. m
,. b....
- r E,?.d. q%.. V
.L y
\\o
!p;.. %' 3?'f.%;.,q,n,.p.:'?.. hl+.<.$y.., %;..
- h.,
. y.,'a
.i
~
yN:i.L'yM,s;pi,'1 o.,#n. p 3
- .b, l..... e};..sh./
4>
s,
v '..
'W 8
J.
s '
c > w. m.. q;.... + ;A h P *i';
Qt,%g,; i,'};:; %.f
- e l'C' *. <',.Y?
- W
.,,,,';6
- ' ' ' + ;f
!l;'.,'.A.,I. Q.xyJ.
- s
- t. !! \\) n 3 ny. S.~M50. m.:.. ;~.:...J.% ry. Q..p.1
. 4. ; '.
- y..;.
- .
- :u....
j
- n
\\
I&;e'\\ '.l ~~4[^ ', j ;..
9Q:5 II.hf.ibcSINbb...;. 56. a.4%....,C.' E:
.D'+_".15..ffM1$T64486
- ! h.\\.,e. n$ N, [ A..
W n.
.s
.. g t
G6/e36
' Ol 2 mm (c) 93% TD Nondensifying (V) As-Fabricated (d) 93% TD Nondensifying (V) As-Irradiated Transverse Section Transverse Section From Rod #48, Local Burnup 'i.50 GWd/mtU p$
'n..s * '.'. s, 1 = '4.'* w.rj
, a m%,- Q.t,pt*. a p ' 1
+7
-.. y- -,,,,,.
/ c 1
m
. C.(r.>
Ns.
4 r:
tny.:..gQ-,,o~Y 1g r,., y e
.-1
.ss
~.
~;
hV1%+p+.n'L&,';'
- n. t. \\
7
'fr,.',
v
$.hh.f..
f^97.eWd%.1
..'l s 4?' i.l
' ;i* r,. 4 e & i' M.
,1 Ly. ( ;;'.{. ;., f; M i.Q yi.,.
ni 5.,.,. /ry.s.,y' 4. 8
,g
%,u h x"',s.A'p,6.,,,.
.(f.
T.,.%
g.%w.,:7m-f. y,q.' if..tgij(,.~.*
yw;f.
~
~ a ~c s
~
ee 3
s.
Mu he-'t.s. n A m:...r g '- *inc30908 is....
'A e
.s.a..
51723 w
e
.i
. 2. c.
...rs.
2_ : _.
- ..Wrd. 4. I L_....
l u.
. M,Y. 3 a. Ed
e.,
'.c
.c c
o Figure 2.
Densification and Swelling Behavior of UO Fuels Irradiated in Calvert Clif fs-I 2
I I
I I
I I
+e -
nienames AGALYass Futt NET FIT fvet e
~
N M StW9MTU e
'I8 p
g I AGO 4 CTCLE A0 Q
' p p *"
w O
'e#
M468w3MTW
[,g g e o
'p e
4 (9
i 2:4:swomte
==
O g
3 a
3cycts see i.
ea asas same antues g
7 __ __ _ _
E u
=
A
.g.3 rat
" 88'EU "ee' rest m testa O :
m
- 3.0
""g["*
E 8 354seweMis oe oa es
- 3.s seessaarvios A es
'E O
D O
o,
-4.0 meesseenvies G 8 as (v) g.
(
l I
l l
l 1
1 I
i.
M t
ESTABATES teltGUP. 0wSMTU l
-