ML20031A141
| ML20031A141 | |
| Person / Time | |
|---|---|
| Site: | Calvert Cliffs  |
| Issue date: | 09/30/1981 |
| From: | ABB COMBUSTION ENGINEERING NUCLEAR FUEL (FORMERLY |
| To: | |
| Shared Package | |
| ML19268A437 | List: |
| References | |
| CEN-183(B)-NP, NUDOCS 8109210039 | |
| Download: ML20031A141 (15) | |
Text
!
,/
,1 i
4 APPLICATION OF CENPD-198 TO ZIRCALOY COMPONENT DIMENSIONAL CHANGES 9-t CEN-183(B)-NP i
SEPTEMBER, 1981.
s w.
[
COMBUSTION ENGINEERING, INC.
WINDSOR, CONNECTICUT l
0109210039 810915 PDR ADOCK 05000317 1
,,..,,.,..,_...._,.,.,.___,,_.,____-,,,,_-_,,,..__,,..,-,,,.__,_.....,-m.,,,,
' i~
i j
/
i
)
LEGAL NOTICE.
THIS REPORT WAS PREPARED AS AN ACCOUNT OF WORK SPONSORED BY COMBUSTION ENGINEERING, INC. NEITHER COMBUSTION ENGINEERING I
NOR ANY PERSON. ACTING ON ITS BEHALF:
,,e A.
MAKES ANY WARRANTY OR R' PRESENTATION, EXPRESS OR J
IMPLIED INCLUDING THE WARRANTIES OF FITNESS FOR A PARTICULAR PURPOSE OR MERCHANTABILITY, WITH RESPECT TO THE ACCURACY, COMPLETENESS, OR USEFULNESS 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 RFSPECT TO THE USE OF,OR FOR DAMAR 6S RESULTING FROM THE USE OF, ANY INFORMATION, APPARATUS, METHOD OR PROCESS DISCLOSED IN THIS REPORT.
4 1
i i
l e
j l
l l
I l
APPLICATION OF CENPD-198 TO ZIRCALOY COMPONENT DIMENSIONAL CHANGES
1.0 INTRODUCTION
This document provides additional information on the analytical model used to predict irradistion-induced dimensional changes of Zircaloy components.
It also provides comparisons of analytical results to actual high-fluence measurements as justification for the acceptability of the analytical.model f
to predict dimensional changes at high fluence levels.
In August,1979, the NRC provided a Staff Evaluation Report (SER) on Topical Report CENPD-198, "In-Reactor Dimensional Changes in Zircaloy-4 Fuel Assemblies" (January,1976), and its supplements (Supplement 1, January,1978, and
~
Supplement 2, November,1978). The SER concluded that CENPD-198 and Supplements 1 and 2 were acceptable for reference in licensing applications with two conditions:
a) A description must be given of the metallurgical status of the components being analyzed (for comparison to those in CENPD-198)
- and, b) The applicability of the analytical model was limited to Zircaloy~
grr.wth strains corresponding to axially averaged fast neutron fluences not exceeding 4x1021 nyt (>0.821 Mev), unless additional data could be supplied for higher exposures.
It is the purpose of this document to describe the current analytical model used to predict dimensional changes, and to provide comparisons of high-fluence dimensional change data to analytical predictions.
It is the conclusion of this document that the analytical model (the SIGREEP Computer Code) is acceptable for use in predicting the irradiation-induced dimensional changes of recrystallization annealed (RXA) Zircaloy-4 guide tubes and stress relief annealed (SRA) Zircaloy-4 fuel rods.
2.0 DESCRIPTION
OF DIMEtJSIONAL CHANGE ANALYSIS
2.1 Background
Topical Report CENPD-198, which was submitted in January, 1976, presented two empirical correlations which had been developed by Combustion Engineering to characterize the permanent dimensional changes that had been observed to occur in the Zircaloy-4 components of nuclear fuel assemblies. Specifically, the first of these correlations was developed using data Trom stress relief f
annealed (SRA) fuel rods either from C-E reactors or of designs similar to those used in C-E reactors; and the second of these correlations was developed using data from the recrystallization annealed (RXA) axial structural components of C-E fuel assemblies or from components metallurgically similar to the C-E fuel assembly structure.
However, the scope of that topical report did not describe the specific criteria, or the methods, by which the two correlations were applied by C-E in the design of fuel assemblies. The criteria and methods were supplied by Supplement 1 to CENPD-198, submitted in January,1978. Supplement 1 described a computerized Monte Carlo technique which formed a resultant joint probability density function from random combinations of variables. The upper or lower 95% probability value
{
was selected from this function, as appropriate, and (after correcting for elastic compression, differential thermal expansion, fabrication tolerances, and creep) was compared to the criterion on assembly growth or fuel rod to upper end fitting shoulder gap.
Additional NRC questions, raised during review of Supplement 1, were answered in Supplement 2, submitted in November, 1978. Also discussed in Supplement 2 was the use of a fabrication tolerance as a variable in the statistical analysis, rather l
l l
than as a single worst case value which was added or subtracted following the statistical portion of the calculation.
The SER on CENPD-198 and Supplements 1 and 2 was issued in August,1979. The SER accepted Combustion Engineering's design methodology as properly addressing the i
dimensional change phenomenon, with the exception of the metallurgical and high-fluence extrapolation conditions identified in Section 1.0 of this document. L
~
2.2 Current Analytical Methodology
}
i The method of caiculating probability density distributions for time-dependert and fluence-dependent dimensions, such as fuel bundle length changes and shoulder gap changes, has been refined to include a time history analysis.
Previously, the dimensional changes of the various components were calculated separately for a gisen point in the operating lifetime, and were then combined using the computerized Monte Carlo technique. That method assumed that each of the factors contributing to dimensional change was independent of the others.
In reality, there are some second-order interdependencies. For example, in the mechanisms leading to fuel bundle length change, irradiation-induced fuel bundle growth causes an additional compression of the upper end fitting springs, thus increasing the compressive load in the guide tubes. 'The higher loads in turn cause an increased compressive creep rate of the guide tubes. Therefore, +he net fuel bundla length change at a given time during operation depends on how the irradiation growth and creep in the specific fuel assembly have combined during operation up to that time. The previous analytical method was to assume a value of compressive creep corresponding to an average fuel assembly, and subtract that value from the irradiation growth predicted at end;of-life (EOL).
The current method of calculating E0!. probability distributions involves following the dimensional changes in fuel bundies throughout their lifetime, thus accounting for the interdependency of parameters. The time history analysis of the fuel bundle and the Monte Carlo sampling technique have been combined into the SIGREEP Computer Code. A typical dimensional change analysis is outlined below:
a) The distribution of uncertainties in the parameters that are handled statistically (including component dimensions, guide tube growth coefficient, guide tube creep coefficient, fuel rod growth coefficient, etc.) are input to the SIGREEP Computer Code. Typically, the uncertainties in the component dimensions are corservatively assumed to have uniform distributions (equal probability of being anywhere within the dimensional tolerance range), while other uncertainties are assigned their appropriate normal distributions.
b) An initial set of randomly selected values for each parameter is chosen by the SIGREEP Computer Cooe. The probability of selecting a given value of a parameter is governed by the distribution of the uncertainty of that particular parameter.
The dimensional values chosen, along with innut dimensional constants, determine the fuel bundle's condition (holddown force, clearances, lengths, etc.) prior to irradiation, i.e., beginning-of-li.fe (BOL).
c) Based on the input operating data (neutron flux, time, temperature, etc.) and the values selected for time-dependent and fluence-dependent parameters, the BOL conditions set by Item 5 are revised incrementally as the SIGREEP Code analyzes the fuel through its period of operation. The operating data are conservatively biased toward unfavorable conditions.
d) When the code reaches the specified operating time, the dimensional changes of the various components are complete. A single value of the desired dimension is the result of the time history calculation.
e) The calculated value is stored and the code returns to Item b and selects a new set of v11ues for the parameters. The new set is used to obtain a new result from the calculations of Items c and d.
The SIGREEP Code continues in this manner until a sufficient number of points (typically 2000 points) have been generated to define a probability histogram of the desired dimension. The resultant
~
histogram represents the statistical variation of the desired dimension which can be attributed to the uncertainties of the input parameters.
f) Finally, a value is chosen from the histogram at the desired probability level for comparison to the appropriate criterion (stated in Supplement 1).
r-c w
y
/
3.0 COMPARISON OF ANALYTICAL PREDICTIONS TO ACTUAL DATA In order to demonstrate that the methodology utilized in the statistical model described in Section 2.2 is conservative for dimensional change predictions at high fluences, comparisons were made between the analytical predictions using the methods described in this report and the available post-irradiation examination (PIE) dimensional change data.
The analytical predictions were based on as many of the actual operating conditions associated with the measured components as possible. The operating conditions included operating temperatures, residence times, fast Tiuxes, and accumulated fluences. Analyses were performed for shoulder gap changes and bundle length changes for various assemblies from Maine' Yankee Cycles 1 and la, and from Calvert Cliffs 1 Cycles 1, 2, 3 and 4.
The analytical predictions for the decrease in shoulder gaps are compared to measured values in Figures 1 thru 4.
Figure 1 shows the comparison of shoulder gap changes versus the 95% probability limit predictions for three Maine Yankee bundles at the end of Cycle 1 (a total of 152 shoulder gap measurements) and four Maine Yankee bundles at the end of Cycle la (a total of 155 shoulder gapmeasurements).
Inspection of Figure 1 indicates that only two of the 307 shoulder gap measurements fall outside the predicted upper 95 or lower 95 limits.
Frequency distributions for the shoulder gap changes measured in three Calvert Cliffs 1 (CCl) Batch B bundles at E0Cl are shown relative to their upper and lower 95% probability predictions in Figure 2.
That figure shows that all of the measured shoulder gap changes in these bundles (BT01, BT03 and 8080) are bounded by the predictions. Measured shoulder gap changes for CCl bundle BT03 are shown in Figure 3 at EOC1, E0C2, E0C3 and E0C4, along with the analytical predictions. The upper and lower 95% predictions bound the measured data and the predicted dependence of gap closure on rod fluence agrees well with the measured data.
Figure 4 compares actual shoulder gap changes of standard Batch D fuel rods to predicted shoulder gap changes for CCl bundles D047 (after one, two, and three cycles of operation) and D048 (after one and three cycles of operation).
Again, the predictions bound the data and the predicted relationship between gap closure and rod fluence agrees well with the measured data. -
D k
Predictions of changes in bundle (guide tube) length are compared to measured value\\
in Figures 5 and 6.
Changes in guioe tube length for Me.ine Yankee bundles at E0C1 (three bundles) and E0Cla (four bundles) are compared to their 95% probability-limit predictions in Figure 5.
That figure shows'that the analytical values envelope the actual length changes and predict the length change dependence on fluence quite
-11, although th ere is one-data point from Cycle la that is.
essentially c lower 95% prediction. Bundle length changes for CC1 bundles with up to four cycles of operation are compared to their 95% probability li2.it predictions, in Figure 6.
That figure shows that, except for bundle BT03 at E0C2 and E0C3, j.
the predictions bound the growth data. The measured growth for BT03 at both E0C2 and E0C3 -is slightly less (.020 inch) than the lower 95% prediction. The predicted incremental growths for BT03 (two, three and four cycle differences) and D047
)
(two and.three cycle differences) agree well with the. actual incremental growths,
, which indicates that the analytical modelling of the bundle length dependence on fluence is performed properly.
Although the analytical model tends to overpredict the bundle length change, the predictions are considered acceptable since:
a) The upper 95% prediction of bundle length changes are conservative, so the use of the analytical prediction for comparison to design,
criteria ensures that there will be no axial interference between the fuel assembly and reactor internals.
b) All the actual data show that the b"ndles are' getting longer tiirough-out their life, so considerations v aere minimum bundle lengths
~
are important (such as fuel bund's engagement with the internals) 4 f,
are covered by BOL conditions.
c) The only other consid9 ration which might be affected by overpredicting the bundle length change is shoulder gap clearance. However, shoulder gap is analyzed separately, and the SIGREEP Code has been shown to adequately predict gap cha.nges.
The comparisons described above, all of which are based on recrystallization annealed (RXA) Zircaloy-4 guide tube material and stress-relief annealed (SRA)
Zircaloy-4 fuel rod material, demonstrate that the analytical model (SIGREEP) successfully predicts the dimensional ~ changes _at high fuel rod fluence levels.
\\
4.0
SUMMARY
l The use of a statistical combination of uncertainties in predicting irradiation-induced dimensional changes has been previously approved up to a rod fluence level of 4x1021 nyt (>.821 Mev), while verification of the conservatism of the method was required for higher fluences when data became available. This document provides comparisons of analytical predictions to high-fluence data, and thereby justifies the statistical method throughout the design lifetime of the f
fuel bundles.
The mechanics of handling the statistical considerations, and of analyzing the fuel bundles through a time history of operation, have been combined in the SIGREEP Computer Code. The code selects a set of initial conditions, adjusts the component dimensions during the bundle's life, and calculates the desired EOL parameter while keeping all the operating conditions' consistent. By analyzing enough sets of randomly selected variables, the probability histogram of the desired dimension is determined. The histogram is used to select a value for comparison to the appropriate criterion.
SIGREEP predictions of shoulder gap change and bundle length change have been compared to actual data, out to axial average fluence levels over the active length of 8.7 x 1021 nyt and 8.8 x 1021 nyt (E >.821 Mev) for fuel bundles and fuel rods, respectively. Those comparisons de:..anstrate the acceptability of the SIGREEP analysis to predict irradiation-induced dimensional chbnges throughout the fuel's lifetime.
l I
t i
i
?
4 i
i FIGURE I 4
i COMPARISON OF SHOULDER GAP CHANGES vs PREDICTIONS FOR M.Y. BUNDLES i
I
' AT EOC1 AND EOC1a l
4 l9
=
4 5
i 4
i E
4 i
f i
i li t
O N
1 e
e
- - - ~ - - -, - -
-c--.
-,. me c,
i i
I FIGURE 2' I
SHOULDER GAP HISTOGRAMS AND UPPER AND LOWER 95 PREDICTIONS 1
FOR THREE CCI BUNDLES AT EOC1
~
i i
1 J
9 1
. j f
0 a
I I
l 6
f 4
f I
l t
b"8 h
g e
a,m.
m,em.'
l
g
..c m.
j c _.,
m-6 am.a.
9 9
FIGURE 3 i i l l' s
COMPARISON OF CALVERT CLIFFS I ASSEMBLY BT03 CHANGE OF SHOULDER GAP VERSUS SIGREEP PREDICTION
- l g
9 6
i i
.i e
t i
\\
r
1, t
FIGURE 4 i i ll5 i
IF, COMPARISON OF CALVERT CLIFFS I BATCH D CHANGE OF SHOULDER GAP
[
VERSUS SIGREEP PREDICTIONS
_.m
+. -
t I
-L r
i e
i 4
I t
I b
i
.1 i
l s
1 8
1 P
I-
.,1 i
L ri E
~~~
u a
. 1
'4
1 m
e T
~
s.
e O
SN O
IT C
I D
E l
R P
t 59 R
EW O
L DNA 59 R
E P
PU sv S
E 5
G E
N R
A U
G H
I C
F H
TGN EL E
B e
U i
G Y.
M F
O
' N O
S I
RA P
MO C
4 i
O FIGURE 6 i
COMPARISON OF CALVERT CLIFFS I BUNDLE LENGTH CHANGES VERSUS SIGREEP PREDICTIONS WWWW iN V
k t
F i
i 4
4 W
N
+
1
.m
,e u
- -