ML18026B015
| ML18026B015 | |
| Person / Time | |
|---|---|
| Site: | Browns Ferry  |
| Issue date: | 05/23/1984 |
| From: | Office of Nuclear Reactor Regulation |
| To: | |
| Shared Package | |
| ML18026B014 | List: |
| References | |
| NUDOCS 8406040120 | |
| Download: ML18026B015 (20) | |
Text
1
~P,S azgI, UNITEDSTATES NUCLEAR REGULATORY COMMISSION
~
WASHINGTON, D. C. 20555
~*"
SAFETY EVALUATION BY THE OFFICE OF NUCLEAR REACTOR REGULATION RELATED TO EVALUATION OF FUEL INITIAL CONDITIOI'IS USED IN TENNESSEE VALLEY AUTHORITY BWR TRANSIENT ANALYSIS BROWNS FERRY NUCLEAR PLANT UNITS 1, 2
AND 3 DOCKET NOS. 50-259/260/296 1. 0 INTRODUCTION In a letter dated January 22, 1982 (Ref. 1), the Tennessee Valley Authority (TVA) described a systalI transient model (TVA-TR81-01, Ref. 2) for the Browns Ferry Nuclear Plant based on the RETRAN-02 code (Ref. 3).
This application requires the use of fuel initial conditions (principally gap conductance) from a separate fuel thermal performance model.
In response to staff questions in this area, the licensee indicated (Ref. 4) that -the COMETHE III-J code (Ref. 5) will be used to provide initial conditions for RETRAN.
At the time of our evaluation of the BWR transient model, the COMETHE code had not been reviewed by the staff and, therefore, did not constitute an acceptable source of gap conductance data.
As a consequence, our approval (Ref.
- 6) of TVA-TR81-01 informed the licensee that approved methods should be used to generate the initial conditions.
The following report deals with the resolution of this condition of approval.
2.0 FUEL INITIAL CONDITIONS OK l PJIAA;
, OlNIL.
~oo cl 0 QIA O '
loH
<A.,
~ore"'OK
.rlA
~
<ma.O.'he use of a fuel pe'rformance code (such as COMETHE) is required in several areas of plant safety analysis.
The applications range fralI establishing the Doppler coefficient contribution to the overall power defect to providing the initial fuel conditions for the loss-of-coolant accident (LOCA) analysis.
For steady-state BWR analysis, the Doppler coefficient contribution is not critical (although a large error in fuel temperature or Doppler may result in unexpected rod position requirements and a Technical Specification violation in predicting l
reactivity).
The LOCA analysis, on the other hand, is used to determine reactor operating limits.
The fuel performance calculations are, therefore, more important in this application and subject to more stringent review.
The application of COMETHE proposed by TVA falls between these two extremes; the "pI code will be used to provide initial conditions for non-LOCA BWR transient analysis.
In BWR transient analysis, the moderator feedback (which controls most events) is strongly dependent on fuel temperature and gap conductance.
t
)I
These two parameters are calculated for both core average and hot channel conditions with the COMETHE code.
3.0 THE COMETHE CODE COMETHE III-J (Ref. 5) is one of a series of COMETHE codes'eveloped by a Belgian firm, Belgonucleaire S.A., and introduced to this country in 1977 by the. S.
M. Stoller Corporation under the auspices of the Electric Power Research Institute (EPRI).
The EPRI LWR Fuel Rod Modeling Code Evaluation Program (Ref.
- 7) investigated six different fuel performance codes (including COMETHE) for potential use by the electric utility industry.
Based on the ability of each code to predict both the thermial and structural response of Zircaloy-clad U02 fuel, COMETHE III-J was judged to be the most versatile and accurate of the codes examined.
Although more recent versions of the COMETHE code are now available (e.g.,
COMETHE III-L, Ref. 8),
EPRI has continued to support the version of the code designated COMETHE III-J.
This work has resulted in severaI additional reports demonstrating the predictive ability of the code (Refs. 9-ll) and recommending modifications (Ref. 12).
However, the modified version of the code exhibits no clear improvement over the original III-J version and neither EPRI nor TVA has adopted these changes.
Additional documentation on COMETHE III-J (and,the code itself) is available under license from EPRI or directly from Belgonucleaire.
We have been awar e for some time of the potential use of the COMETHE code in plant safety analysis applications.
Our dealings in the past on this subject (Refs.
13-17) have been with the Electric Power Research Institute, which does not hold a reactor operating license.
Prior to TVA's proposed use of COMETHE, we were aware of only a single case in which a version of COMETHE was cited in a licensing action.
In that case (Ref.
18)., the Connecticut Yankee Atomic Power Company responded to our request for information on high burnup fission gas release with results of calculations using the COMETHE III-H code.
Recog-nizing the fact that the code had not yet received staff review and approval,
and that the licensee had previously responded (Ref.
- 19) to our request with approved Westinghouse methodology, the COMETHE calculations were largely ignored.
The TVA proposal, therefore, constitutes the first formal request for using the COMETHE code in plant safety analysis.
4.0 APPLICATION OF COMETHE AT BROWNS FERRY Tne reactor core in the TVA BWR transient model is represented in one (axial) dimension by twelve active (fueled) and two reflector (nonfueled) core volumes.
The resulting model, shown in Figure 2-4 of TVA-TR81-01 (Ref. 2), also considers a core bypass volume.
The moderator density and fuel temperature for each active volume are used to provide the neutronic feedback in the calculation.
A hot channel model, using twenty-four active fuel and two unheated volumes, is utilized in the critical power ratio (CPR) calculation.
The control volumes and heat conductors used by RETRAN to model the core are simple approximations of the actual fuel loading.
In order to obtain the one-dimensional kinetics parameters (e.g., macroscopic cross sections),
the results of another code are used to assure that RETRAN's neutronics solution is consistent with a multi-dimensional calculation at the start of the transient.
This second code is a three-dimensional, steady-state simulator called CORE and is described in TVA-MDS-553 along with procedures for collapsing the cross-.sections to one dimension.
The CORE model for fuel temperature utilizes a one-dimensional, cylindrical fuel rod temperature solution which can be applied at different axial and radial locations.
The cladding surface temperature and heat flux are applied as boundary conditions with the fuel-to-cladding gap conductance used as an input variable (from COMETHE).
The RETRAN model itself utilizes a low fuel rod gap conductance that is uniform axially and constant during the transient calculation.
This approach (of using a constant, average value of gap conductance) has been previously approved in other applications.
Several averaging techniques (e.g.,
constant
- 4 or weighted average) have also been approved in other applications.
Our evaluation (Ref. 6) of RETRAN found the TVA methods also acceptable in this regard.
We further noted that high (i.e., biased) values of gap conductance were utilized for the hot chanhel calculation, as opposed to the low values used in the core average calculation.
Because the sense of the conservatism reverses between these two applications, we found this approach acceptable as well.
Because L CPR increases with increasing hot channel gap conductance, the maximum value expected over the exposure range of interest is used in the licensing analysis.
The licensee has pr ovided (in Response No.
8 of Reference
- 4) generic hot channel gap conductance values for the major fuel designs used at Browns Ferry (i.e.,
GE 8xS, GE 8xSR and GE PSxSR).
These values are representative of an assembly continuously operated at the maximum average planar linear heat generation rate (MAPLHGR) limit.
Our own calculations (Ref. 20) confirm that, over a reasonable range of burnups, thi s is an appropriate method of determining maximum values of gap conductance.
The core-averaged value of gap conductance used in the BWR transient analysis is evaluated for each reload core and state (i.e., burnup) analyzed.
The licensee has provided a representative value for this calculation (for a full core of GE P8xSR fuel at 105 percent power at end of cycle) as well as the maximum bounding values of core-averaged gap conductance versus core-averaged burnup (Ref. 21).
Both the core-averaged and hot channel gap conductance values are further discussed by the licensee in Reference 22.
We have examined these values a>>d note that they are representative of gap conduct-ances predictea by GEGAP-III (Ref. 23),
an approved fuel performance code utilized by the Browns Ferry fuel vendor (General Electric) in previous reload safety analyses.
5.0 EVALUATION A number of considerations went into determiningthe scope of this review.
First, the application of COt1ETHE in the Browns Ferry safety analysis is
1'imited.
The code will be only used to provide fuel initial conditions in the BWR transient analysis and the licensee will continue to rely on the fuel vendor for fuel design, LOCA (i.e.,
MAPLHGR) and other analyses.
- Second, the licensee has shown (Ref. 4) that significdnt changes in the gap conductance values used in the BWR transient analysis have only a limited effect on the calculated'PR.
- Third, a reasonable amount of demonstrative (as opposed to descriptive) literature (Refs.
7, 9-12) already. exists in the public domain on COMETHE III-J independent of that submitted by TVA.
Although these documents are not sufficient for a detailed technical review of COMETHE (e.g., it is not possible to determine whether the code conforms with Regulatory Guide l. 126 - Ref. 24), they are generally sufficient to assess the predictive capability of the code.
Fourth, it has been determined (Ref.
- 25) that the licensee must not only demonstrate the adequacy of the code itself, but also a proficiency in using it.
On the basis of these considerations, we have elected not to perform a focal technical review of the COMETHE III-J code.
Rather, this evaluation will:
1.
Oetermine that the version of the COMETHE III-J code proposed for use by the Tennessee Valley Authority and the version qualified in the open literature (Refs.
7, 9-12) are one and the same.
- 2. Confirm the adequacy of COMETHE III-J for gap conductance (as inferred from fuel temperature) predictions with more recent experimental data.
3.
Oemonstrate the licensee's proficiency in using the code for plant safety analysis.
To assure that the TVA version of COMETHE III-J is identical to the official EPRI release
As a result, the licensee has confirmed (Ref. 22) that the current TYA version is identical to the version offered by EPRI.
To confirm the adequacy of COMETHE III-J with more recent experimental
- data, the previously-reported work (Refs.
7, 9-12) was re-examined.
We found that these reports are not all complimentary.
For example, References 9 and 10 indicate that COMETHE III-J is a poor (non-conservative) predictor of fission gas release, particularly from unpressurized (BWR) rods with high measured.
release fractions.
It should be noted, however, that the application proposed by TVA is not very sensitive to errors in predicted fission gas release (as opposed to rod internal pressure and LOCA analysis).
Boiling water reactor transient analysis, however, is sensitive to fuel-to-cladding gap conductance, which is inferred from experimentally-measured fuel temperature data.
Adaitional data of this kind were sought to verify the predictive ability of COMETHE.
It was proposed that several rods from Instrumented Fuel Assembly IFA-432 (Refs. 26-29) would serve to confirm the predictive capability of the COMETHE code.
This test consisted of several short-length fuel rods with centerline therwocoupl es, built-in pressure transducers and other instrumentation.
Fuel dimensions were approximately the same as the GE 8x8 fuel used at Browns Ferry.
The assembly was irradiated in the Heavy Boiling Water Reactor (HBWR) at Halden, Norway over a wide range of power levels, including those antici-pated at the Browns Ferry facility.
This test was sponsored by the Nuclear Regulatory Commission and operated by Battelle Pacific Northwest Laboratory.
We consider this, and several other recent Halden tests, to be state-of-the-art irradiation experiments.
The licensee provided (Ref. 22) the COMETHE III-J predictions of fuel center-line temperature (and input assumptions) for Rods No. I, 2 and 3 of IFA-432.
These rods are of equivalent design (i.e., General Electric unpressurized 8x8 fuel) with the exception of the initial fuel-to-cladding gap size.
The as-fabricated diametral gap size for Rod I (9 mils),
Rod 2 (15 mils), and Rod 3 (3 mils) may be contrasted with the nominal values for General Electric's 7x7 (12 mils) and 8x8 (9 mils) fuel designs.
The results of this compar ison between measured and predicted data are shown in Figure I with Rod 1, 2 and 3
data represented by their respective numerals.
Although the scatter in the
TVRr'CGlIETHE PREDICT IGNIS GF IFA-'f32 S
~ O
~O e ~
CSI li li 2
2 2'2 22 2
2
(
1 2,, r l
1 1
1 (U
C ej C$
0 S
0 0
S Q
OO O'J aOO OO Q3 1
1 1
3 3
3 33 3
2
~3 3
'53
3 3
3
.3 3
3 600 800 1000 1200 1900 measured Centerline Temperature, (C) 1600 1aoo I
Figure 1.
COHETHE-predicted fuel centerline temperature as a
function of measured fuel centerline temperature.
data is large, the overall trend follows a best-estimate (45') line.
A statistical analysis of the IFA-432 data, as shown in Table I, has also been performed.
This information is similar to that provided by the. licensee in Table Y of Reference 22 except that the real, rather than the absolute, value of the difference between measured and predicted temperature is used.
We have also included the same statistical information in terms of relative error (i.e., the real difference between measured and predicted temperature divided by the measured temperature).
The latter form accounts for the low I
uncertainty at low power.
That is, at zero power, the fuel temperature (with very little uncertainty) should be equal to the coolant temperature.
In either case, the results show that COMETHE III-J is, on the average, slightly non-conservative for the data shown (by approximately 18'C or 2>).
The corresponding standard deviation (115'C or 9.3%) is scmewhat smaller than
- expected, based on the NRC's own fuel performance codes (155'C, Ref. 30).
However, the licensee has included a canparison between COMETHE III-J and an older NRC code called GAPCON-THERMAL-3 (Refs.
31-32) relative to the IFA-432 data.
The results suggest that COMETHE provides better predictions of the experimental data than the NRC code.
Further examination of the predictive ability of COMETHE III-J is shown as a function of gap size
( Figure 2), exposure
( Figure 3) and local power level
( Figure 4).
Although the code appear s to exhibit-a distinct non-conservative bias with respect to decreasing gap size (and perhaps with respect to increasing power level as well), the results are reasonable when restricted to gap sizes (and power levels) expected in the Browns Ferry analysis.
We conclude that COMETHE III-J will provide reasonable gap conductance predictions (as inferred from fuel temperature data) for the range of conditions expected in the Browns Ferry application.
We further conclude that the conservati sms employed in performing these calculations (e.g.,
hot channel analysis at the MAPLHGR limit) are sufficient to justify the use of COMETHE III-J for the application sought by TYA (non-LOCA BWR transient analysis)
The licensee's analysis of IFA-432 data would ordinarily be sufficient to demonstrate a proficiency in using COMETHE III-J for plant safety analysis as
TABLE 1
SUMMARY
OF IFA-432 COMETHE RESULTS Y=(T -T )/T Rod Number Thermocoup1e Location Number of Points 432-1 Upper 21
-24. 38'C 47.95'C
-0. 0156
- 0. 0335 432-1 Lower 37
+61. 97'C
- 31. 50'C
+0. 0508 0.0281 432-1 Both 58
+30. 71;C 56.44'C
+0.0268 0.0439 432-2 Upper 432-2 Lower 36
+98. 69'C 79.01'C
+0. 0698 0.0584 432-2 Both 36
+98. 69'C
- 79. 01'C
+0. 0698 0.0584 432-3 Upper 35
-180.49'C 63.24'C
-0.1466
- 0. 0508, 432-3 Lower 37
-56.16'C 37.93'C
-0.0626
- 0. 0421 432-3 Both 72
-116.60'C 81.00'C
-0.1034
- 0. 0626 ALL RODS 166
-18.44'C 115.30'C
-0. 0204 0.0930
, TVA,'"COi'lETHE PREDICTIONS OF IFFY-R3'
~
n
(
ri Rod 2
sJ LLJ uJ Rod 1
fg I
Q
<</~l I
~ L
- ll Qp C~
C I
Roil J LA I
ClE 8 <8 GE 7>7 004
.Ol
.015 Ini". '
1 Di @met.ra 1
Gap Size, (Inches)
.02 Figure 2.
Mean error
$ n predicted fuel centerllne temperature as a function of 1nit1al gap size.
, T'<)Fir'CGI'1ETHE PREDICT I GNS GF IFA-'t32 S
~J 1 >l W)
Q I
0 GJ Ul lg Ql 0S Ul lU8 I
Ql U
U$
CL O
LA C'0 I
3 3
333 333 3
3 1
f 3
1 3 3$
3 3
3 3
33 33 3
3 33-3 3
3 3
3 2
2 2
1 2
1 1
1 1
1 11 11 1
1 222 11 1
333
'3 33.
3 333 33 100
. 200 Cu~u 1 a t ive Exposure, (Days)
F)gure 3.
Norma11zed error )n predicted fuel centerline temperature as a function of exposure.
,TVR~'COIIETHE PRED ICTI ONS GF IFFl-932 S
L.
f0 S
cL Ul C l S
O GJ Ql Q
UIL Ul fg8 I0 Ql 0
TJS Q-C LD I
y3 1
3 3
3333 3b 3y y 39 33 3 9 3 3
3 3
3 3
2 2
2 2g 9
2 2 2 2 I
)2g 12 2
1 1
)
11 22 1
1 1
2 3
3 3
3 3
3 3 3 9 3 1
3 3 3 3
3 9
3 3 3 3
8 10 12 Local Power Level, (kM~ft )
Figure 4.
Normal ized err or in predicted fuel centerl one temperature cssnr04nh n+
1 nrel nnwar 1 wel.
required by Generic Letter 83-11 (Ref. 25).
Shortly after the verification data (IFA-432) were agreed
- upon, however, it came to our attention that IFA-432 predictions with COMETHE III-J were also being published (Ref.
11) by a third party,.Science Applications, Inc., under the sponsorship of the Electric Power Research Institute.
It is not possible to determine what impact the EPRI report may have had on the TVA submittal (published some 5 months later).
The EPRI report, which includes data in addition to IFA-432, shows the same predictive trends for IFA-432 as shown by the licensee's submittal.
Because both reports rely on the same code (COMETHE III-J), this is to be expected.
Based upon other details in the TVA submittal (which do not appear in the EPRI report), it is our judgement that the licensee's report was developed autonomously; We conclude, therefore, that the licensee has demonstrated a level of proficiency in using this code that justifies application in plant safety analysis.
We regard the EPRI report (EPRI NP-2992) as an additional demonstration of COMETHE III-J's pre-dictive ability and a confirmation of the overall trends already shown for this code by the licensee.
6.O CONCLUSIONS We have examined the'roposed application of COMETHE III-J in the BWR transient analysis of the Browns Ferry Nuclear Plant.
We find this application acceptable':
Because of the manner in which conservatisms are employed in this application, the manner in which gap conductance affects the analysis, and the license qualification requirements of Generic Letter 83-11, we restrict this approval to analyses performed by TVA for TVA BWR facilities of limiting pressurization transients (non-LOCA) in reload evaluations as descrihed in TVA TR 81-01.
Principal Contributor:
Dated:
May 23, 1984
REFERENCES 1.
L. M. Mills (TVA) letter to H.
R.
Denton (NRC) dated January 22, 1982.
2.
S.
L. Forkner et al.,
"BWR Transient Analysis Model Utilizing the RETRAH Program,"
Tennessee Valley Authority Topical Report TVA-TR81-01, December 31, 1981.
Attachment to Reference 1 above.
3.
"RETRAN A Program for Transient Thermal-Hydraulic Analysis of Complex Fluid Flow Systems," Electric Power Research Institute,'Report EPRI NP-1850-CCM, Volumes I, II and III, May 1981.
4.
L. M. Mills (TVA) letter to H.
R.
Denton (NRC) dated January 20, 1983.
5.
P.
Verbeek and N.
- Hoppe, "COMETHE III-J:
A Computer Code for Predicting Mechanical and Thermal Behavior of a Fuel Pin," Part 1: General Description (Nonproprietary),
Belgonucleair e S.
A. (25 rue du Champ de Mars, 1050, Brussels, Belgium) Report BN7609-01, September 1976.
6.
D.
B. Vassallo (NRC) letter to H.
G. Parris (TVA) dated April 7, 1983.
7.
H.
R.
Freeburn et al. (Combustion Engineering, Inc.), "Light Water Reactor Fuel Rod Modeling Code Evaluation," Electric Power Research Report EPRI NP-369, March 1977.
8.
N.
Hoppe et al.,
"COMETHE III-L:
General Description," Belgonucleaire Report BN 8201-04, January 1982.
9.
G. Leppert et al. (Science Applications, Inc.), "Analysis of Fission Gas Release Measuraoents Using the COMETHE IIIJ and FCODE-Alpha Computer Codes," Electric Power Research Institute Report EPRI NP-1954, July 1981.
10.
S.
Lee et al. (Science Applications, Inc.),
"Comparison of COMETHE-IIIJ and FCODE-BETA Fission Gas Release Predictions With Measurements,"
Electric Power Research Institute Report EPRI NP-2903, March 1983.
11.
L.
Rayes et al.
(Science Applications, Inc.),
"COMETHE-IIIJ Predictions of the Fuel Centerline Temperature in the IFA-431, IFA-432, and IFA-513 Assemblies," Electric Power Research Institute Report EPRI 'NP-2992, May 1983.
12.
K. Lindquist et al. (S.
M. Stoller Corporation),
"Evaluation and Modification: of COMETHE III-J," Electric Power Research Institute Report EPRI,NP-2911, Mar ch 1983.
13.
R.
Lobel (NRC) letter to F.
E.
Gelhaus (EPRI) dated April 26, 1977.
14.
J.
C. Voglewede (NRC) letter to F.
E.
Gelhaus (EPRI) dated September 8,
1977.
15.
F.
E. Gelhaus (EPRI) letter to J.
C. Voglewede (NRC) dated November 23, 1977.
16.
F.
E. Gelhaus (EPRI) letter to J.
C. Voglewede (NRC) dated June 20, 1978.
17.
D.
G.
Eisenhut (NRC) letter to F.
E.
Gelhaus (EPRI) dated September 25, 1978.
18.
D.
C. Switzer (CYAPCO) letter to A. Schwencer (NRC) on "Haddam Neck Plant Fission Gas Release" dated fiovember 8, 1977.
19.
D.
C. Switzer (CYAPCO) letter to A. Schwencer (NRC) on "Haddam Neck Plant Fission Gas Release" dated January 4, 1977.
20.
D. L. Acey and J.
C. Voglewede, "A Comparative Analysis of LWR Fuel Designs,"
U.S.
Nuclear Regulatory Canmission Report NUREG-0559, July 1980.
21.
R.
Rogers (TVA) communication with R. Clark (NRC) dated June 7, 1983.
22.
J.
K.
Lee and J.
F. Burrow, "Validation of COMETHE III-J for Gap Conductance Calculations,"
Tennessee Valley Authority Report EAS-138 (October 25, 1983).
Transmitted by L. M. Mills (TVA) letter to H.
R.
Denton (NRC) dated November 21, 1983.
23.
"GEGAP-III:
A Model for the Prediction of Pellet-Cladding Thermal Conductance in BWR Fuel Rods," General Electric Company Report NEDC-Z0181 (Proprietary),
November 1973.
24.
"An Acceptable Model and Related Statistical Methods for the. Analysis of Fuel Densification,"
U. S.
Nuclear Regulatory Cawoi ssion Regulatory Guide
- 1. 126, Revision 1, March 1978.
25.
D.
G. Eisenhut (NRC), Generic Letter No. 83-11 to All Operating Reactor Licensees on "Licensee gualification for Performing Safety Analyses in Support of Licensing Actions" dated September 26, 1983.
26.
C.
R.
Harm et al., "Test Design, Precharacteri zation and Fuel Assembly Fabrication for Instrumented Fuel Assemblies IFA-431 and IFA-432,"
Battelle Pacific Northwest Laboratory Report NUREG/CR-0332 (PNL-1988),
November 1977.
27.
C.
R.
Harm et al.,
"Data Report for the NRC/PNL Halden Assembly IFA-432,"
Battelle Pacific Northwest Laboratory Report NUREG/CR-0560 (PNL-Z673),
April 1978.
28.
E.
R. Bradley et al.,
"Data Report for the NRC/PNL Halden Assembly IFA-432:
April 1978 - May 1980, "Battelle Pacific Northwest Laboratory Report NUREG/CR-1950 (PNL-3709), April 1981.
29.
E.
R. Bradley et al., "Final Data Report for the Instrumented Fuel Assembly (IFA)-432," Battelle Pacific Northwest Laboratory Report NUREG/CR-2567 (PNL-4240),
June 1982.
30.
E.
T. Laats,'.
Chambers and N.
L. Hampton, "Independent Assessment of the Steady State Fuel Rod Analysis Code FRAPCON-2,"
EG&G Idaho, Inc.
Report EGG-CAAP-5335, January 1981.
31.
D.
D. Lanning et al.,
"GAPCON-THERMAL-3 Code Description," Battelle Pacific Northwest Laboratory Report PNL-2434, January 1978.
32.
D.
D. Lanning et al.,
"GAPCON-THERMAL-3 Verification and Comparison to In-Reactor Data," Battelle Pacific Northwest Laboratory Report PNL-2435, September 1978.