ML20210N733
| ML20210N733 | |
| Person / Time | |
|---|---|
| Site: | Calvert Cliffs  |
| Issue date: | 02/04/1987 |
| From: | Office of Nuclear Reactor Regulation |
| To: | |
| Shared Package | |
| ML20210N699 | List: |
| References | |
| NUDOCS 8702130199 | |
| Download: ML20210N733 (9) | |
Text
,
- p uog%
UNITED STATES NUCLEAR RECULATORY COMMISSION o
{
- )
WASHINGTON, D. C. 20006
\\,e... /
SAFETY EVALUATION BY THE OFFICE OF NUCLEAR REACTOR REGULATION BALTIMORE GAS AND ELECTRIC COMPANY CALVERT CLIFFS NUCLEAR POWER PLANT, UNIT NOS. 1 AND 2 DOCKET NOS. 50-317 AND 50-318 REVIEW OF TOPICAL REFORT CEN-151(B)-P SUPPLEMENT 1-P
" IMPROVEMENTS TO FUEL EVALUATION MODEL"
1.0 INTRODUCTION
The FATES 3 code was developed by Combustion Engineering (C-E) and submitted to the NRC for review in July 1981 (Ref. 1).
The NRC completed the safety evaluation and approved (Ref. 2) the FATES 3 code for C-E safety analyses,.
but imposed a restriction on the grain size used in fission gas release calculations.
The FATES 3 version with the NRC-imposed grain size restriction
(
has been referred to as FATES 3A by C-E.
The new FATES 3 code version, recently developed by C-E (Ref. 3) and submitted by Baltimore Gas & Electric Company (BG&E) (Ref. 4) to NRC for review and approval, is referred to as FATES 38.
The FATES 3B code is identical to the FATES 3 and FATES 3A versions of the code with modifications only to the fission gas release and thermal expansion models.
The purpose of this technical evaluation is to determine if these modifications are appropriate for the code's intended appifcations.
As a result of this review, NRC has submitted questions (Ref. 5) to BG&E to clarify the modifications to the code and the intended use of the code.
BG&E and C-E have responded to these questions (Ref. 6) and they will be referred to extensively in this evaluation.
2.0 APPLICATIONS OF FATES 3B IN LICENSING ANALYSES The NRC has requested (Ref. 5) that BG&E define the licensing analyses to which the FATES 3B code will be applied.
BG&E and C-E have replied (Ref. 6) that FATES 3B will be used for all future BG&E licensing applications.
The FATES 3B licensing applications will remain the same as those defined for FATES 3 (Ref. 7),
BG&E and C-E have also stated that the input will remain the same as those defined in Reference 7 for FATES 3.
3.0 FISSION GAS RELEASE MODEL MODIFICATIONS The basic structure and mechanisms for fission gas release remain the same between FATES 3B and FATES 3.
Both codes have two basic modes of release:
- 1) steady state gas release that is dependent on burnup, fuel temperature and grain size; and 2) a short term release model that is related to grain growth, which, in turn, is fuel temperature and timo dependent (via grain 0702130199 070204 PDR ADOCK 05000317 P
.g.
boundary sweeping).
The differences between the two codes' steady-state release models are that the FATES 3B model has a stronger burnup dependence at high burnups, the grain size dependence is reduced and it is no longer dependent on dynamic grain size, only on the fabricated grain size.
The FATES 3B short-term release model assumes that grain boundary sweeping is the primary mechanism for release during power and fuel temperature inc^reases on the order of minutes-to-hours.
The actual mechanism for
' release during power increases is still not well understood, even though a considerable amount of research has been done.
However, this issue is somewhat academic, as long as the FATES 3B model adequately predicts the release during these power increases.
The FATES 3B grain growth model has been based on unirradiated isothermal grain growth data, as well as ir-radiated data.
In addition, this model has also been adjusted to agree with the power bumping data at both low and high burnups.
4.0 THERMAL EXPANSION MODEL MODIFICATIONS The differenc'e between the FATES 3 and FATES 3B thermal expansion models
(
is relatively small.
The FATES 3 (and the FATES 3A) model assume a fixed inner diameter for thermal expansion of an annular pellet.
This model only allows the fuel pellet to expand at diameters greater than the fixed diameter and thus, does not allow expansion at or below this fixed diameter.
This model is common in accounting for expansion in a solid cylinder.
The FATES 3B model retains the same U0 thermal expansion properties as FATES 3,butallowsforfreethermalekpansionofanentireannularpellet (i.e., no fixed inner diameter), which is more consistent with the. nature of a cracked annular fuel pellet that exists in an instrumented fuel rod.
This improvement allows for a more realistic prediction of fuel temperatures for the fuel rods with centerline thermocouple, particularly at low linear heat generation rates (LHGR's).
5.0 EVALUATION OF MODIFICATIONS TO FATES 38 The purpose of this section is to evaluate the modifications to the fission gas release and thermal expansion models within the FATES 3B code.
As noted earlier, the purpose of this evaluation is to determine if the modifications employed in FATES 3B are appropriate for the code's intended applications.
With this objective in mind, the evaluation of FATES 3B has consisted of two approaches:
1)
The code predictions have been verified against well-characterized fission gas release and thermal data from fuel rods that have operated in the ranges of the code's intended application.
2)
Audit comparisons of FATES 3B predictions of typical licensing analyses have been performed with an independent NRC fuel performance code.
t
. The verification of FATES 3B predictions against well-characterized data is considered to be of greatest importance, while the code audit comparisons are used to confirm the observations made from the comparisons against the data.
The evaluation presented below is divided into fission gas and thermal
. predictions.
5.1 Evaluation of the FATES 3B Fisson Gas Release Predictions The FATES 3B code has been verified against fission gas release data from 92 fuel rods that have operated under both steady-state power and bumped power (i.e., the rod power is increased from a steady-state value to a higher power level and held for a given amount of time) operating conditions.
The rod powers of this fission gas release data are typically near the maximum powers allowed by the Technical Specifications for C-E fuel.
This is because fission gas release at lower powers is very low, and thus, is of little concern.
The data presented by C-E also represents a wide range of burnup, between 7 mwd /kgM to 62 mwd /kgM (rod average).
This data more than adequately covers the range of operating conditions intended for C-E fuel rods.
g The FATES 3B predictions, along with the pertinent operating conditions of the 92 fuel rods with measured gas release values, are presented in Tables 3-1 through 3-12 of Reference 3.
In addition, Figure 3-1 of this reference provides a comparison of FATES 3B predicted versus measured release values for the 92 fuol rods.
The predicted-minus-measured release values versus burnup are provided in Figure 3-2 to illustrate that there are no biases in the FATES 3B predictions with burnup.
However, an initial review of the predicte'd versus measured comparison of release values, provided in Figure 3-1, has indicated that the code may be underpredicting fission gas release when measured values are greater than 25% release (absolute).
A closer examina-tion of this comparison to data has shown that the code underpredicts 19 out of the 24 rods, with measured values greater than 25% release with a mean underprediction of 6% release (absolute).
A question of bias in the FATES 3B code predictions was presented to BG&E by the NRC (Ref. 5).
Combustion Engineering has responded (Ref. 6) that the bias exists in three of the data sets used for the FATES 3B comparisons and, because there is no bias in the prediction of the remaining data sets, there is no significant bias in the code.
Combustion Engineering has offered an experimental reason for the bias in two of the data sets (consists of 10 rods with measured release values above 25%) that were bumped in power in the Petten Reactor.
The explanation for the bias is related to the unique way that Petten has bumped these rods in power.
The staff's consultant has examined this evidence (Ref. 8) for the experimental bias in these rods and agrees with C-E that the bias is due to the power bumping apparatus used for these rods in the Petten Reactor and this does not represent
.4 -
a bias in the FATES 3B prediction.
In addition, independent evidence of the bias in this data has been provided from the predictions of a proprietary fuel performance code available to the NRC.
This proprietary code has been verified against a large amount of fission gas release data and has shown a slightly greater tendency to underpredict this data than the FATES 3B code.
This helps to confirm the conclusion of a bias in these two data sets.
Combustion Engineering has offered no experimental reason for the bias in the third data set, that consists of three ramped rods.
However, C-E has noted that the NRC-approved FATES 3A code also underpredicted this data. The staff's consultant has also examined several reports, both proprietary and publicly available, and found that these three rods have been used in the verification of several fuel performance codes. All of these codes have significantly underpredicted the gas release in these three rods.
Consequently, the staff's consultant has concluded that it is likely that the fission gas release data from these three rods is also biased, due to unknown experimental reasons.
The NRC has requested (Ref. 5) that a representative FATES 3B licensing calcu-lation of end-of-life rod pressures be provided to NRC for audit comparisons against NRC fuel performance codes.
Combustion Engineering has provided the
(
results, along with the input, of the requested calculation.
The code input of the fuel rod fabrication parameters for this calculation has previously been provided in Reference 7.
The basis for the power history input for this calculation is very important and has also been previously described in References 7 and 9.
NRC approval for the use of this input for end-of-life rod pressure calcula-tions has been provided in References 2 and 10 for normal and extended burnup operation, respectively.
In brief, the fabrication and power history inputs are based on conservative (i.e., upper bound or lower bound, where appropriate) values that will achieve the maximum possible fission gas release for the C-E fuel design in question.
The same input used for the FATES 3B calculation has been input into the NRC audit code GT2R2 (Ref. 11).
The options chosen for the GT2R2 calculations were selected to provide a "best estimate" prediction of fission gas release up to extended burnup.
This included the selection of the revised ANS 5.4 gas release model (Ref. 12) in GT2R2 (Ref. 11).
It should be noted that using these options in GT2R2 does not assure a real-istic estimate of fission gas release and end-of-life rod pressures, because the code has been shown to be somewhat conservative, even with the "best estimate options" (Ref. 11).
Consequently, this ensures a small, but conser-vative, bias to the GT2R2 predictions.
Also, because the revised ANS 5.4 release model has relatively slow release kinetics, the short term power increases in the C-E power history have been artificially increased in time to allow an equilibrium release value to be calculated for these power increases.
This approach will bound the amount of release possible from these short term power increases and, thus, is conservative for this calculation.
0
, The comparison of the GT2R2 and FATES 3B predictions are proprietary and, thus, will only be summarized in this report.
The GT2R2 predictions of internal rod pressure were generally lower or equal to the FATES 3B predictions within about 70 psi at high burnups (greater than 45 mwd /kgM rod average).
Because the fission gas release predictions for the GT2R2 code have been shown to be slightly conservative it is assumed the rod pressure calculation is conserva-tive.
This indicates that the FATES 3B code may be slightly conservative in the prediction of internal rod pressures.
This is somewhat puzzling, since the FATES 3B fission gas release predictions were found to be best estimate in this evaluation.
A possible explanation for the similar predictions of internal rod pressure between GT2R2 and FATES 38, with small differences in fission gas release predictions, may be the way each code handles the internal void volumes in the fuel rod.
If FATES 3B handles the effects of the void volume decrease in the rod with burnup more conservatively than the GT2R2 code, this would explain the similar rod pressure predictions of these two codes.
From the above evaluation, it is concluded that the FATES 3B prediction of internal rod pressures is either best estimate or slightly conservative.
Fuel
(
temperature and, thus, fuel rod power is the primary driving force for fission gas release.
Therefore, the use of bounding power histories is the major source of conservatism in the C-E calculation of end-of-life rod pressures.
The bounding nature of these power histories, along with conservative rod fabrication input, assures that the rod pressures calculated with the FATES 3B code will also be bounding.
Consequently, it is concluded that the use of the FATES 3B code for calculating end-of-life rod pressures is acceptable for licen-sing applications if the fuel fabrication and power history input described in References 7 and 9 and approved by NRC in References 2 and 10, respectively, are used.
5.2 Evaluation of the FATES 3B Thermal Predictions The FATES 3B code has been compared against a large volume of centerline fuel temperature measurements from fuel rods instrumented with centerline thermo-couples. This data was also submitted in the FATES 3 submittals (Refs. I and
- 13) to NRC for verification of this code's thermal predictions; however, the FATES 3B submittal has corrected the thermocouple data for decalibration with neutron fluence that was not known about nor corrected in the earlier submittal.
The data comes from a well-characterized set of instrumented rods, with a wide range of fill gas pressures, that were irradiated in the Halden Test Reactor.
In Figures 3-3 through 3-5 of Reference 3, C-E has provided the following FATES 3B comparisons with data; predicted versus measured, predicted-minus-measured versus burnup, and predicted-minus-measured versus LHGR, respectively.
These figures have shown a consistent overprediction in relation to the corrected thermocouple data.
The conservative bias appears to remain rela-tively constant with measured temperature, burnup and LHGR, with a few excep-tions.
The conservatism appears to be less at very low LHGRs (less than 4 KW/ft) or very low burnups (less than 0.5 mwd /kgM).
These operating ranges are not critical for licensing applications.
A statistical measure of the
. thermal conservatism of the FATES 3B predictions has been provided by C-E from a statistical analysis of the predicted-minus-measured data.
The statistical analysis is very similar to the NRC and C-E analyses performed in References 2 and 13, respectively, in support of the FATES 3 submittal.
Each of these analyses has calculated a mean bias and standard deviation from the code's predictions of the thermocouple data.
The standard deviation and mean bias
- of the FATES 3B predictions will be compared against those calculated earlier by.C-E (Ref. 13) and NRC (Ref. 2) for FATES 3 to determine the difference in relative uncertainty and conservatism between these two codes.
The evaluation of the standard deviation calculated by C-E for FATES 3B is presented below, followed by an evaluation of the mean bias in the code's predictions.
The standard deviation calculated by C-E for FATES 3B (Ref. 3) is nearly iden-tical to that calculated by C-E for FATES 3 (Ref.13); however, it is less than that calculated by the NRC (Ref. 2) for FATES 3.
The reason for the lower standard deviation calculated.by C-E for both the FATES 3 and FATES 3B analyses is because they have eliminated three rods from their statistical analysis.
that were included in the earlier analysis by the NRC.
Combustion Engineering has stated that the elimination of one of the Halden test rods was justified because both lower and upper thermocouples from this rod were erratic, k
indicating failure.
Combustion Engineering eliminated the other two rods because they were a typical of the C-E design and the C-E code significantly overpredicted centerline temperatures and fission gas release for these two rods.
The staff's consultant has reviewed this data and agrees that it is very likely that the thermocouples had failed in the former rod, and that the design and strong thermal feedback calculated for the latter two rods are atypical of the C-E designs.
Therefore, the standard deviation C-E has calculated for the FATES 3B code is found to be representative of their fuel design (i.e., prepressurized fuel rods) and has not changed significantly from the standard deviation calculated for FATES 3.
The mean bias in centerline temperature predictions calculated by C-E for FATES 3B (Ref. 3) is smaller than that calculated by NRC (Ref. 2) and C-E (Ref. 13) for FATES 3.
Both the NRC (Ref. 2) and C-E (Ref. 13) calculated mean l
biases for FATES 3 were nearly identical, which means the elimination of the erratic data by C-E has reduced the standard deviation, but has not changed the mean bias of the code. The reduction in the bias for FATES 3B compared to FATES 3 appears to be due to the improved thermal expansion model and, to a l
lesser extent, the correction of the measured thermocouple data for decali-bration.
A reduction in the conservative bias of the code's centerline temperature predictions, is acceptable if the code maintains an acceptable margin of conservatism for its intended licensing application.
The most important of the thermal licensing applications for the code is in its use for LOCA initialization. A very good discussion of the direct relationship between centerline temperature and stored energy is provided in the NRC Safety i
Evaluation Report of FATES 3 (Ref. 2).
The staff's consultant has used the same statistical approach to evaluate whether or not there is adequate conservatism in the FATES 3B predictions of centerline temperature and, thus, stored energy.
l l
l
1 The centerline margin used in the FATES 3 review for LOCA initialization was that the code conservatively predict fuel temperatures with a 95% probability at a 95% confidence level.
This conservative margin is not a regulatory requirement, but has been used in several previous code reviews.
Therefore, based on this premise, the code uncertainty limit is:
1.8 times the standard deviation where 1.8 is the appropriate statistical factor, given the 95%
probability at the 95% confidence level and the number of data points in the data set. The standard deviation used is the standard deviation of the FATES 3B predictions of the data set discussed earlier.
There are two primary contributions of conservatism in the C-E calculation of stored energy:
- 1) The FATES 3B conservative bias that has been the focus of this review, and 2) the conservatism introduced by input.
The C-E input to the FATES 3B calculation for LOCA initialization has several different conser-vative parameters that provide a significant, proprietary, conservative margin on predicted centerline temperature and, thus, stored energy.
Combustion Engineering has stated (Ref. 6) that the margin provided by the input to FATES 3B is on the order of that found for the FATES 3 code in Referent.e 13.
From examination of these conservative margins, the staff's consultant has determined that the mean bias in the FATES 3B prediction of centerline temper-ature is less than the uncertainty limit determined above, i.e., the mean bias in the centerline temperature is less than 1.8 times the standard deviation, was the case for the FATES 3 code.
However, it has been found that the code predictive bias plus the input bias is much greater than the uncertainty limit.
Consequently, it is concluded that the thermal bias in the FATES 3B code is adequate for its intended application for determining LOCA initial conditions.
Combustion Engineering has provided the results of a representative licensing calculation for LOCA initialization with the FATES 3B code, similar to the calculation for end-of-life rod pressures, discussed earlier in Section 5.1.
The rod fabrication and power history inputs have also been provided.
It should be noted that the short term powers to which the code is ramped are close to, but not equal to, the Technical Specification LHGR limit allowed by the C-E design. The initial conditions from FATES 3B are then input into the C-E emergency core cooling system (ECCS) code STRIKIN-II.
The STRIKIN-II code is then ramped up to the Technical Specification LHGR limit for the LOCA analysis, as described in References 2 and 7.
The same input used for the representative FATES 3B calculation for LOCA initialization has been input into the NRC audit code GT2R2 (Ref. 11).
Two different options for fuel relocation were selected to provide conservative and a "best estimate" prediction of fuel centerline temperatures early-in-life, when both the GT2R2 and FATES 3B codes predict the maximum temperatures.
A comparison of the GT2R2 audit results to the FATES 3B results confirm the conservative bias in the latter code for early-in-life predictions.
A spe-cific value for the conservative bias in the FATES 3B code was not possible from this audit comparison, but a range of bias in the code was estimated and this range bounded the C-E calculated bias.
For general information, it is noted that the later-in-life FATES 3B thermal predictions were closer to the "best estimate" GT2R2 predictions.
. From the above evaluation, the staff concludes that the FATES 3B thermal predictions are adequate for the codes intended applications.
6.0 CONCLUSION
S The staff has completed the evaluation of the proposed modifications to the Combustion Engineering fuel performance code, designated as FATES 38, submitted by Baltimore Gas & Electric Company.
The modifications to the fission gas release and fuel thermal expansion models have been found to be acceptable for use in licensing applications submitted by Baltimore Gas & Electric Company.
The approval of the fission gas release and fuel thermal expansion models is based on; 1) verification of FATES 3B predictions against well-characterized fission gas release and thermal data that has operated in the ranges of the code's intended applications, 2) verification that the code and input are adequately conservative for its intended applications, and 3) audit comparisons by the staff's consultant using an independent NRC fuel performance code to verify the FATES 3B results of typical licensing analyses.
- 7. 0 REFERENbES 1.
" Improvements to Fuel Evaluation Model," CEN-161(B), July 1981 (Proprietary), Combustion Engineering, Inc.
2.
Letter from Robert A. Clark (NRC) to A. E. Lundvall (BG&E), " Safety Evaluation of CEN-161 (FATES 3)," March 1983.
3.
" Improvements to Fuel Evaluation Model," CEN-161(B)-P, Supplement 1-P, April 1986 (Proprietary), Combustion Engineering, Inc.
4.
Letter, J. A. Tiernan (BG&E) to Ashok C. Thadani (NRC), May 9, 1986 (letter transmitted Reference 3 to NRC).
5.
Letter, S. A. McNeil (NRC) to J. A. Tiernan (BG&E), September 29, 1986.
6.
" Response to NRC Questions on FATES 3B," CEN-345(B)-P, October 17, 1986, Combustion Engineering, Inc.
7.
" Partial Response to NRC Questions on CEN-161(B), ' Improvements to Fuel Evaluation Model,'" CEN-193(B), Supplement 2-P, March 21, 1982 (Proprietary), Combustion Engineering, Inc.
8.
J. C. LaVake and M. Gaertner, "High Burnup PWR Ramp Test Program, Topical Report: Background Ramp Test Results," 00E-ET/34030-4, CEN-402, December 1982.
9.
" Extended Burnup Operation of Combustion Engineering PWR Fuel,"
CENPD-269-P, Revision 1-P, July 1984, Combustion Engineering, Inc.
. 10.
Letter, E. J. Butcher (NRC) to A. E. Lundvall (BG&E), " Safety Evaluation Report for Extended Burnup Operation of Combustion Engineering PWR Fuel," CENPD-269-P, Revision 1-P, October 10, 1985.
11.
M. E. Cunningham and C. E. Beyer, "GT2R2: An Updated Version of GADCON-THERMAL-2," NUREG/CR-3907 (PNL-4178), September 1984.
12.
Letter, C. E. Beyer (PNL) to J. C. Voglewede (NRC), " Revised ANS 5.4 Model," May 24, 1982.
13.
" Supplemental Information on FATES 3 Stored Energy Conservatism,"
CEN-220(B), October 5, 1982 (Proprietary), Combustion Engineering, Inc.
Date: February 4, 1987 Principal Contributor:
D. Fieno I
e l
<