ML20235X262
| ML20235X262 | |
| Person / Time | |
|---|---|
| Site: | Grand Gulf  |
| Issue date: | 02/28/1989 |
| From: | Eastburn M, Patchana T, Smith F SYSTEM ENERGY RESOURCES, INC. |
| To: | |
| Shared Package | |
| ML20235X241 | List: |
| References | |
| RPAS-SR-89-007, RPAS-SR-89-7, NUDOCS 8903130395 | |
| Download: ML20235X262 (29) | |
Text
_ - _ _ - _ .- _
Attachment 1 to AECN89/0037 Criticality Safety Analysis of the Grand Gulf Nuclear Station, Unit.1 Spent Fuel Storage Racks with Gaps in the Neutron Absorbing Panels RPAS-SR-89/007 February, 1989 .
Author: - M M. R. Eajtburn, Senior Nuclear, Engineer Author: !A M T. Patchana, Seni r Nuclear Engineer Author: - -
F. H. Smith, Lead Nuclear Engineer Review: . amt -
R. B. Lang,Ct.ea Nuclear Engineer Approve: [ M C. B. Franklin, Manager Reactor Physics Analysis Nuclear Engineering Services Department MSU System Services Inc.
I l
l g993130395 890227 pga ADOCK0500g6 )
P
. i
'.6 i- . .
t TABLE OF CONTENTS
. EXECUTIVE
SUMMARY
. . . . . . . . . . . . . . . . . . . . . .. 1 l INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . 2 l
METHODOLOGY . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 r
l RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 REFERENCES . . . . . .. . . . . . . . . . . . . . . . . . . . 18 APPENDIX A . . . . . . . . . . . . . . . . . . . . . . . . . . 19
LIST OF TABLES Table'1: Gap Size'and Frequency Distributions . . . . . . . . . 10 Table 2: Analyzed Boraflex. Gap Probability Distribution . . . . 11 Table 3: Summary of Mechanical Uncertainties . . . . . . . . . 12 Table 4:' Assembly Design Parameters . . . . . . . . . . . . . . 13 Table 5: Rack Design Parameters . . . . . . . . . . . . . . . . 14 Table 6: Calculation of 95/95 K-eff . . . . . . . . . . . . . . 16 Tabis 7: 95/95 K-EFF Response Matrix . . . . . . . . . . . . . 17 WP 1
9 LIST OF TIGURES Figure 1: 1/8 Assembly Map of ANF 9 x 9 LTA . . . . . . . . . . 9 I
i
)
J l
l 1
l L-__-___-_-______-_ _ _ _
l -. EXECUTIVE
SUMMARY
This report describes a criticality safety analysis of the. Grand-Gulf Nuclear Station, Unit 1 (GGNS-1) High Density Spent Fuel l Storage Racks (HDSFSR) fully loaded with the most reactive GGNS-1 fuel. This analysis was performed to determine the effect of potentia {deteriorationoftheneutronabsorbingmaterial, Boraflex
, Gaps in Boraflex panels have been found in HDSFSR manufactured by-the Joseph oat Corporation. This analysis uses in-service blackness test data and Boraflex coupon test data to estimate
" worst-case" distributions for gap size and gap frequency, hereafter referred to as the' analyzed distributions. Criticality calculations were performed for various combinations of Boraflex gap sizes and gapped panels per cell. From each calculation, a 95
% probability /.95 % confidence level upper limit for the k-effective of a given Boraflex. gap configuration was obtained. The results of these calculations were statistically combined based upon the probability of each Boraflex gap configuration. This resulted in a overall 95/95 upper limit of 0.9496 k-effective for GGNS-1 HDSFSR's should the analyzed Boraflex gap distributions be realized.
Since the neutron multiplication factor is conservatively calculated to be less than the-NRC acceptance criteria of 0.95, this analysis confirms that the GGNS-1 HDSFSRs can safely accommodate the maximum reactivity GGNS-1 Cycle 1 through Cycle 4 fuel with the presence of gaps in the Boraflex panels.
1 Boraflex is a trade name for a boron carbide dispersion in an elastomeric silicone matrix manufactured by Bisco Products, Inc.
(BISCO).
a
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - - - - - - - - - - - . _ _ _ _ _ _ _ _ . - - - - - _ _ _ _ _ _ _ _ _ . - - - - - a
INTRODUCTION The spent fuel storage racks in the Grand Gulf Nuclear Station were originally designed and qualified for fuel assemblies of 3.5%
enrichment. Since then, the racks have been re-qualified for fuel with a maximum distributed average enrichment of 3.61 wt. % U-235 for ANF supplied 8x8 fuel and 3.47 wt. % U-235 for ANF 9x9-5 fuel.
Both analyses were performed under very conservative assumptions and with a large margin below the NRC limiting multiplication factor of 0.95 [1,2].
In September of 1987, the USNRC sent Information Notice No. 87-43 (3) to all operating licensees, alerting the recipients that gaps had been found in the Boraflex spent fuel rack poison panels at i Quad Cities Unit 1. In response to this issue, blackness testing was performed on the GGNS-1 HDSFSR Boraflex panels by neutron measurements. The tests located a number of gaps in the irradiated storage cells, most of which were less than one inch in size. The maximum gap size was 1.4 1 0.2 inches. It is important to note that only twenty five percent of the storage cells were observed to contain gaps in as many as two Boraflex panels in the same axial plane, only one cell had three co-planar gaps and none had four gaps in the same plane.
The present analysis has been performed to demonstrate the safeh*
criticality of the racks, even if fully loaded with GGNS-1 Cycle 4 ANF 9x9-5 LTA fuel, the most reactive GGNS-1 fuel bundle. The multiplication factor of the rack has been calculated using estimated " worst-case" distributions of Boraflex panel gap size and gap frequency. The distributions analyzed were based on the GGNS-1 HDSFSR blackness testing measurements, similar measurements performed at Quad Cities Unit 1 and other industry data reported in Reference 4. The Quad Cities data is of particular interest since its HDSFSR, also manufactured by J. Oat Corp., is virtually identical to the GGNS-1 rack. The maximum gap size is modeled as 6.0 inches, baced upon the EPRI Boraflex shrinkage correlation reported in Reference 4. This is significantly larger than the l current maximum in-service HDSFSR gap size of 4.0 inches reported i by Quad Cities.
l 1
1 l
METHODOLOGY Maximum Expected Boraflex Gap Distribution l'
l The results of the GGNS-1 HDSFSR Blackness testing were evaluated i to determine the applicability of the EPRI Boraflex shrinkage correlation, which related the Boraflex panel shrinkage to gamma fluence [4). The high gamma fluence portion of the EPRI correlation is based primarily on in-reactor Boraflex coupon tests which were conducted in a significantly harsher environment than is present in the GGNS-1 HDSFSRs. This environment includes the presence of a strong neutron field, high intensity /high energy gamma field and high temperature operating environment.
As expected, the calculated gamma fluence for the Boraflex panels in the GGNS-1 test area was bounded by the EPRI correlation.
Since there is insufficient data at this time to provide a basis for a GGNS-1 specific correlation, EPRI's predicted maximum shrinkage of four percent was used to establish a maximum gap size. All of the Boreflex shrinkage is assumed to result in a single gap corresponding to approximately four percent of the panel's length. This assumption is quite conservative, since some of the shrinkage will occur in areas of the panel where a gap has not formed. Four percent of the panel length corresponds to a gap size of 5.75 inches, which was rounded up to 6.0 inches in the analysis.
GGNS-1 and Quad Cities data indicate that the size and location of Boraflex gaps are nearly randomly distributed throughout the rack. Since criticality is sensitive to the gap size and axial location as well as the number of panels with gaps, an estimated
" worst-case" distribution was developed. This distribution was established by multiplying a conservatively established gap size distribution by a conservatively established gap frequency distribution. These distributions are based upon the measurements at GGNS-1 and Quad Cities summarized in Table 1.
The distribution of the gaps per cell was based upon the number of gaps within in each cell, independent of axial location. This is a very conservative approach since it results in approximately 15 percent of the cells having 4 panels with gaps when none were observed in the GGNS-1 data. Similarly, approximately 12 percent of the cells are estimated to have 3 panels with gaps when only 1 has actually been observed. In order to provide additional conservatism, the probability for 3 gaps per cell was increased to be equal to the probability of 4 gaps per cell. The analyzed distribution was compared to Quad Cities' data (Table 1) and was found to be more conservative. The final results are presented in Table 1.
-l I
1Nhr GGNS-l' analyzed gap sizeidistribution was established by normalizing the GGNS-1 m3asured distribution to a gap size of 6.0
-inches. This distribution was conservatively adjusted by assuming all gaps of less than 2 inches were 2-inches. In order to provide additional conservatism, the probability of 3 and 4 inch gaps sizes was increased. This results in a very conservative distribution relative to both the GGNS-1 and Quad Cities measurements. The final distribution is presented in.
Table'l. The combination of the gap size distribution and the ,
frequency distribution is presented in Table 2.
Analysis Methodology and Model AMPX-KENO was used in'the criticality analysis of the spent fuel storage racks. Originally developed by the Oak Ridge National Laboratory for criticality safety analyses, AMPX uses the'123-group SCALE cross-section library and the NITAWL routine to derive' weighted cross-sections for U-238 in the resonance region with the Nordheim resonance integral treatment. SCALE is an acronym for Standard Computer Analysis for Licensing Evaluations.
Output from NITAWL is supplied to the KENO program, a three-dimensional' Monte Carlo neutron tracking code that calculates the system multiplication factor, k-effective. The SCALE code system is' described in Reference 5.
Even though the AMPEX/ KENO methodology has been extensively
- benchmarked by the nuclear industry, critical experiment ,
benchmarks were performed using SSI specific methodology in order '
to determine the calculational uncertainty associated with SSI specific applications. The results and' application of this benchmark are described in Appendix A.
The reactivity effects of manufacturing tolerances were previously determined using CASMO as describad in Reference 2 and summarized in Table 3. The model assumptions used in the current analysis are consistent with those used'in Reference 2 and therefore the results of the mechanical tolerance analysis are applicable to the current analysis.
The AMPEX/ KENO methodology was applied to the basic rack model (see below) to evaluate several combinations of Boraflex gap I sizes and number of gapped panels per cell. The calculational uncertainty, described in Appendix A, and mechanical tolerance uncertainties were added to the results of each case to provide a 95/95 upper limit k-effective for each Boraflex gap configuration. Additional k-effective values were determined by interpolation. From these calculations a response matrix which relates a specific Boraflex gap configuration to an, upper limit k-effective was derived (see Table 7 of the Results section).
l In order to determine the k-effective for a rack with the analyzed gap distribution, 10,000 random samples were taken from l the distribution of Table 2. For each sample the corresponding k-effective was determined from the response matrix. The average l
]
1 L -
c - _ -- _-
j k' effective and the 95/95 uncertainty for the sampling method Were then calculated. The uncertainty was combined with a one-sided 95/95 tolerance factor and added to the average k-offective to obtain a 95/95 upper limit k-effective for a rack with the analyzed distribution of Boraflex gaps.
Fuel Assembly Design Model The most reactive fuel bundle for GGNS-1 is the ANF supplied Cycle 4 9x9-5 LTA. This bundle consists of an array of fuel rods containing UO in zircalloy cladding. The rods have a 138 inch 2
enriched zone with an average enrichment of 3.47 wt. % U-235 and a 6 inch natural uranium reflector at each end. The fuel is modeled as 150 inches of enriched uranium using the as-built distribution of enrichments indicated in Figure 1. This conservatively models the 6.0 inch natural uranium blankets as enriched fuel.
The stack height density is calculated using pellet dishing factors only: chipping and stacking factors are not used. The number densities of Gd 0 3 bearing fuel are calculated as if the gadolinia were removed All ziralloy, in both the cladding and the fuel channel is modeled as Zr-2. The rack and assembly structural material above and below the active fuel is replaced with a thirty centimeter water reflector.
Design parameters are summarized in Table 4.
Storage Rack Model The KENO storage rack model consists of a single half-height cell of the storage' rack with fuel and channel, reflected at the mid-plane and on all four vertical faces to form a radially infinite, axially symmetric array. A fixed neutron absorber material (Boraflex), 0.070 inches thick, is positioned between the stainless steel walls of each cell. This arrangement provides a nominal center-to-center cell spacing of 6.2585 inches. The 0.063 thick inch stainless steel box which defines the fuel assembly storage cell has an typical inside dimension of 6.0 inches. The rack parameters are shown in Table 5.
In the analysis, Boraflex was modeled as pure B,C with B-10, B-11 and carbon densities set at the 95/95 lower limit of the vendor's
! assay data. These densities, in gm/ce, are 0.119, 0.526 and 0.179 for B-10, B-11 and carbon, respectively. The B-10 areal f
densjty,basedona0.063-inchBoraflexthickness, is 0.0190 The size of the Boraflex gap is varied by replacing g/cm .
i Boraflex with water at the mid-plane. The geometric modeling of I k
Boraflex gap configurations, such as 2 opposing gaps vs 2 !
adjacent gaps was investigated and found to be conservative. j 1
I l l
1 i
L_________ }
9 ABNORMAL AND ACCIDENT CONDITIONS l Reactivity Effect of Temperature and Void References 1 and 6 demonstrated that the temperature and void coefficients of reactivity are g negative for the rack. Since a l conservative temperature of 20 C was used as the design basis, the temperatures actually expected in the pool will result in lower k-effective values.
Other Abnormal / Accident Conditions Other abnormal and/or accident conditions evaluated in Reference 6 remain valid. These include (1) positioning an assembly outside the storage rack, (2) eccentric asemmbly location within the storage cells, (3) dropped fuel assembly, and (4) lateral movement of a rack module, all of which have been found to have a negligible or negative reactivity effect. Similarly, removal of the zirconium flow channel has been shown to have a negative reactivity effect.
CODES AND STANDARDS Applicable codes, standards, and regulations, or pertinent sections thereof, include the following:
General Design Criterion 62, Prevention of Criticality in Fuel Storage and Handling.
USNRC Standard Review Plan, NUREG-800, Section 9.1.1, New Fuel Storage, and Section 9.1.2, Spent Fuel Storage.
USNRC letter of April 14, 1978, to all Power Reactor Licensees - OT Position for Review and Acceptance of Spent Fuel Storage and Handling Applications.
USNRC Regulatory Guide 1.13, Spent Fuel Storage Facility Design Basis, Rev. 2 (proposed), December, 1981 USNRC Regulatory Guide 3.4, Nuclear Criticality Safety in Operations with Fissionable Materials at Fuels and Materials Facilities, Rev. 2,, March, 1986.
ANSI /ANS-57.2-1983, Design Requirements for Light Water Reactor Spent Fuel Storage Facilities at Nuclear Power Plants.
i ANSI N210-1976, Design Objectives for Light Water Reactor Spent Fuel Storage Facilities at Nuclear Power Plants, i
l l
I l
,ANS-8.17-1984, Criticality Safety Criteria for the Handling, Storage and Transportation of LWR Fuel outside Reactors.
LIST OF CONSERVATIVISMS i To ensure that the true reactivity will always be less than the -
calculated reactivity, the following conservative assumptions are made in the model: j The moderator is gure, unborated water at a conservative temperature of 20 c. The temperatures actually expected in the pool will result in lower k-effective values.
The lattice of storage cells is assumed infinite in the
- radial direction, i.e. , no credit is taken for radial neutron leakage (except in the assessment of certain abnormal / accident conditions).
Neutron absorption in minor structural members is neglected, i.e., spacer grids are replaced by water. A water reflector of 30 cm replaces the rack and fuel assembly structural material above and below the active fuel.
Every fuel assembly is assumed to be unirradiated and of the 9x9-5 LTA type, with 3.47 wt. % U235 average enrichment. This fuel has the highest reactivity based upon in-core cold lattice k-infinity calculations.
1 Enriched fuel is modeled to the top of the active fuel length in the analysis, replacing the natural uranium blanket.
The fuel density was conservatively established. l No credit is taken for the neutron absorption of burnable poisons. The typical reload fuel at Grand Gulf contains Gadolina burnable poisons.
All gaps are assumed to occur at the axial elevation (the midplane) which maximizes the reactivity effect.
No credit is taken for increases in Boraflex B-10 density in due to shrinkage of the Boraflex material.
Within credible limits, the values of the Boraflex width, length, and composition were modeled for maximum reactivity.
l l
The maximum gap size modeled is 6.0 inches, larger than {
the expected 5.75 inch maximum based on four percent shrinkage as predicted by EPRI (4),-and larger than the 4.0 inch maximum gap measured to date.
Gap size measurements of panels with multiple gaps were treated as a panel with one large gap when developing the gap size probability distribution. The reactivity effect of one large gap in a panel would be expected to be more than that of two small gaps.
For the purpose of calculating reactivity effects, a gap falling within a gap size interval is assumed to be at the upper limit of that interval.
All gaps in a cell are assumed to be at the same level, maximizing the reactivity effect due to the interaction of the gaps.
- _ _ _ _ _ _ _ _ _ _ . a
1 l
1' Figure 1: 1/8 Assembly Map of ANF 9 x 9 LTA 1'
L1 ML1 M1 MH1 MH1 l ---__-----------_-----------------
l MLl* MH1 H2 M2*
MH1 H2 W l
H2 H2 .j i
l SCR-Rod Type Wt. % U-235 Wt. % Gd 23 0 **
L1 2.32 0.0 Large ML1 2.90 0.0 Large MLl* 2.90 5.0 Large M1 3.20 0.0 Large M2* 3.20 5.0 Small MH1 3.48 0.O Large H2 4.13 0.0 Small W Water Rod SCR Spacer Capture Rod J
r i l-l l
l I
Table 1: Gap Size and Frequency Distributions l
l PROBABILITY Measured Analyzed i
PANELS / CELL GGNS QC-1 GGNS O 0.27 0.38 0.22 1 0.19 0.21 0.20 2 0.27 0.27 0.28 3 0.12 0.10 0.15 4 0.15 0.04 0.15 PROBABILITY Measured 1 Analyzed GAP SIZE (IN) GONS QC-1 GGNS O-2 0.710 0.79 0.38 2-3 0.110 0.14 0.36 3-4 0.065 0.07 0.12 4-5 0.065 0.08 ,
5-6 0.050 0.06 {
1)- GGNS-1 measurements normalized to 6 inches l
l l
l l
l
4 Table 2: Analyzed Boraflex Gap Probability Distribution PANELS / CELL 0 1 2 3 4 GAP SIZE (IN) 0-2 0.0836 0.0760 0.1064 0.0570 0.0570 2-3 0.0792 0.0720 0.1008 0.0540 0.0540 3-4 0.0264 0.0240 0.0336 0.0180 0.0180 4-5 0.0176 0.0160 0.0224 0.0120 0.0120 5-6 0.0132 0.0120 0.0168 0.0090 0.0090
~.
I I
l 1
1 I
I 1
I l
l l
l
l l
Table 3: Summary of Mechanical Uncertainties
-i
's {
r SOURCE OF UNCERTAINTIES delta-k ,
Enrichment Tolerance 0.00340 j
Pellet Density 0.00219 Pellet Dish 0.00042 Pellet Diameter 0.00038 Boraflex Thickness 0.0065 Stainless Steel Thickness 0.00158 Cell Pitch 0.00506 Channel Bulge 0.00462
=
Total Uncertainty (RMS Sum) 0.01041 (See Reference 2) + t-t, I
4 Table 4: . Assembly Design Parameters PARAMETER. SPECIFICATION MODEL VALUE Rod Pitch (in) 0.563 0.~563 Enriched Fuel Length (in) 138.0 150.0 Pellet Density (%TD). 94.5 i 1.5 ,
94.5 Pellet Dish (vol%, no Gd 0 )
23 1.4 1 0.3 1.4 Pellet Dish (vol%, w/ Gd 0 )
23 .0 1 0.3 1.0 Enrichment:
L1 Rod-(% U-235) 2.32 1 0.05 2.32 ML1 Rod (% U-235). 2.90 1 0.05 2.90 ML1 -Rod (% U-235, No Gd) 2.90 1 0.05 2.90 M1 Rod (% U-235) 3.20 i O.05 3.20 M2 Rod.(% U-235, No Gd) 3.20 1 0.05 3.20 MH1 Rod (% U-235) 3.48 1 0.05 3.48 H2 Rod (% U-235). 4.13 1 0.05 4.13 Pellet Diameter:
Type 1 (in) 0.3745 1 0.0005 0.3745 Type 2l(in) 0.3525 i O.0005 0.3525 Clad ID/OD:
Type 1 (in) 0.381/0.443 0.381/0.443 Type 2 (in) 0.359/0.417 0.359/0.417 Water Rod (in) 0.522/0.546 0.522/0.546 Spacer Capture Rod (in) 0.506/0.546 0.506/0.546 k
i l
l l
Table 5: Rack Design Parameters PARAMETER SPECIFICATION MODEL VALUE Cell Pitch (in) 6.2585 1 0.062 6.2585 Boraflex:
Thickness (in) 0.070 1 0.007 0.070 Length (in) 144 1 0.25 143.75 Minimum gm B-10/cm 2 b
Stainless Steel:
1 Thickness (in) 0.063 1 0.006 0.063 1
The 0.006 inch tolerance was changed to 0.0035 inch following the analysis in Reference 2. The 0.006 inch effect is conservative.
l
l RESULTS
'The' KENO k-effective'results for each' gap configuration are provided in Table 6. The KENO uncertainty was statistically combined with the methodology uncertainty and is reported under column "SIGT" in Table 6. This. result is then-multiplied by the appropriate reliability factor and to this
. product the methodology bias and the mechanical tolerance uncertainty are added. The resulting total correction is provided under the column " Total corr." in Table 6. When the total correction is added to the KENO k-effective the 95/95 upper limit k-effective for each Boraflex gap configuration is obtained as indicated in Table 6 and 7. This process is described in detail in Appendix A.
A' response matrix was assembled from the 95/95 upper limit k-effectives of the various gap configurations with intermediate points in the matrix being determined by interpolation. The interpolated values were examined for consistency with the calculated results. The response matrix is shown in Table 7.
In order to determine the system k-effective for a rack with a distribution of gaps, the analyzed gap distribution was sampled 10,000 times and the corresponding k-effective for each Boraflex gap configuration was determined from the response matrix. The system k-effective was determined from the average of these trials. The uncertainty in the system k-effective about this average was also determined. This
. produced a k-effective of. 0.9374 + 0.00729. Applying a one-sided reliability factor results In a a 95/95 upper limit k-eff of 0.9496 for the rack.
This analysis confirms that the maximum reactivity of the expected maximum distribution of gap sizes and gapped panels per cell will be less than 0.95, including uncertainties, with the racks fully loaded with GGNS-1 Cycle 4 9x9-5 LTA fuel and flooded with clean unborated water at the temperature within the operating range corresponding to the highest reactivity.
l 82507074157 05440888516 5f 29792647983 9f 63841284396
/e 33334443468 5 - 99999999999 9K . . . . . . . . . . .
00000000000 82517074157 11350025792 l . 85824848963 ar 19090707799 tr 21212121111 oo 00000000000 ,
TC . . . . . . . . . . .
00000000000 br 50050000000 ao 66656060060 it 00000000000 f l c . . . . . . . . . . .
f ea 22222222222 e RF K )
- A 5 '14668287634 9 54827082836 x
/ T 87361810087 i 5 G 43434243333 d 9 I 00000000000 n S 00000000000 e f . . . . . . . . . . . p o 00000000000 p A
n o . e i d e t t 19473346281 s a S . 86351719077 (
l v 43434242333 u Oe 00000000000 .
c ND 00000000000 d l E . . . . . . . . . . . e a K 00000000000 s C u 6
- f 0 6 6 o. 0 1 n 00000505505 .
e .e 55555151151 0 -
l oG 1 11 1 b N a .
T 6
. 00000000000 0 94190863824 0 f 34968809910 .
f 44750486507 0 e 11112221256
- 99999999999 <
K . . . . . . . . . . .
_ 00000000000 T S G LL I
_ EL S NE 03412341234
_ AC s
_ P/ e m
_ i
_ P . 00000000000 t AN . . . . . . . . . . .
_ GI 02244446666 .
F E
S 12345678901 R A 11
- C l
D; i ..
V <
1 .
l l Table 7:- 95/95 K-EFF Response Matrix PANEL / CELL 0 1 2 3 4 GAP SIZE (IN) 0-2 0.9340 0.9340 0.9340 0.9340 0.9387 2-3 0.9340 0.9344 0.9357 0.9361 0.9400 3-4 0.9340 0.9348 0.9412 0.9427 'O.9485 4-5 0.9340 0.9348 0.9430 0.9540 0.9647 5-6 0.9340 'O.9348 0.9440 0.9698 0.9864 Interpolated values 4
REFERENCES
- 1. S. E. Turner, " Criticality Safety Ana Fuel in the Grand Gulf Nuclear Station,1ysis Spent of Fuel Cycle 4 Racks,"
HI-88274, October 1988,
- 2. " Criticality Safety Analysis for the Grand Gulf Spent Fuel Storage Racks with Cycle 4 9X9-5 LTA (3.47% Average Enrichment," ANF-88-170(P), November 1988.
- 3. FRC Information Notice No. 87-43: " Gaps in Neutron-Absorbing Material in High-Density Spent Fuel Storage Racks," September 8, 1987, SSINS No.: 6835.
- 4. "An Assessment of Boraflex Performar4ce in Spent-Nuclear-Fuel Storage Racks," EPRI Report NP-6159, December 1988.
- 5. " SCALE-2 A Modular Code System for Performing Standardized Computer Analysis for Licensing Evaluation," Oak Ridge National Laboratory, Doc. No.
CCC-450.
- 6. " Licensing Report on High-Density Spent Fuel Racks for Grand Gulf Nuclear Station Unit 1," NRC Docket No. 50-416 Joseph Oat Corp., August, 1983.
l l
l
)
e l
l l
l l
APPENDIX A BENCHMARK CALCULATIONS
- l. INTRODUCTION AND
SUMMARY
This benchmark verifies the NITAWL/ KENO [1] methodology with the 123-group SCALE cross-section library for use in criticality calculations. The NITAWL/ KENO calculations have been benchmarked against critical experiments which have geometries and materials similar to those found in fuel storage racks. Results of these benchmark calculations are consistent with calculations reported in the literature and with the requirements of Reference 2.
- 2. METHODOLOGY Seven critical experiments performed by Babcock and Wilcox (B&W) and fourteen performed by Battelle Pacific Northwest Labs (PNL) were modeled. All criticals were water moderated arrays of UO2 rods in aluminum clad. Most of the criticals contained either stainless steel or aluminum plates. Sone of the stainless steel plates and all of the aluminum plates contained boron. The B&W criticals, described in Reference 3, contained soluble boron in the moderator. The B&W fuel contained 2.46 wt. % U235 while seven of the PNL critical experiments contained 2.35 wt. % U235 fuel and the remaining seven contained 4.29 wt. % U235 fuel. The PNL experiments are described in References 4 and 5.
The lower enriched PNL criticals and the similarly enriched B&W experiments showed virtually identical results even though the geometries were quite dissimilar. The higher enriched PNL criticals, although geometrically similar to the lower enriched PNL criticals, exhibited a markedly lower k-eff. No dependence of k-eff bias on pin pitch, cluster separation or boron, either in solution or in borated plates, could be firmly established. However, an enrichment dependent bias was indicated. A linear least squares fit of the data was used to obtain the behavior of k-effective with enrichment. This results in an increase of the KENO k-effective when fuel with an enrichment greater than about 3.06 weight percent U235 is analyzed. Compared to the bias that would be obtained if a mean were calculated for the NITAWL/ KENO results as a whole, the enrichment dependent bias is a conservative approach for the cases of interest.
- 3. RESULTS The bias uncertainty due to the benchmark methodology was calculated using the methodology of Reference 6. This approach has also been used for a benchmark done by Yankee Atomic [7].
Then, Bias = 0.00326793*ENR - 0.00997 l
l - - _ _ _ - _ _ _ _ - _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ _. _ _ _ _ _ _ - _ _ _ ._ - _ _ _ - - _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ - _ _ _ _ _ _ - _ _ - .
Where ENR is in percent U235 enrichment.
Using Criterion 2 of Reference 6, a benchmark methodology variance of 0.40E-06 was calculated. This variance is combined with the model uncertainties and the KENO standard deviation to obtain the total calculational uncertainty. After the application of an appropriate one-sided 95/95 tolerance factor, the result is summed with the bias and the KENO eigenvalue to obtain the 95/95 critical eigenvalue of the rack (95% probability at a 95% confidence factor). As an added conservatism and to maintain consistency with the benchmark analysis, the following limits are imposed:
- 1. The one-sided tolerance factor times the RMS sum of the method standard deviation (the square root of the method variance) and the KENO standard deviation is to be at least 0.006.
- 2. The enrichment dependent bias is to be equal to or greater than zero.
The total amount of bias and uncertainty added to the KENO results for 4.29 weight percent U235 enriched fuel to obtain a 95%/95%
critical eigenvalue for the benchmark cases is one percent delta k.
l
- p. DV 288154263006152807523 835189401914082395895 222322233233323322222 TE 000000000000000000000 SD 000000000000000000000 000000000000000000000 )
S S
B 777978797326837098181 (
537465462755671657732
- FC 228021891390341567270 l FL 009100990090009999990 e EA 009000390090009999990 e KC . . . . . . . . . . . . . . . . . . . . t s
110111001101110000001 s
s e
100010200000000000000 l FS 000000000000000000000 n FA 000000000000000000000 i EE 000000000000000000000 a KM . . . . . . . . . . . . . . . . . . . . . t s
- S 111111111111111111111 d
N e O LLLL t I AAAA SSSS SSSS a T BBBB SSSS SSSS r A
C T EE01540E005522E005522 d L A NN36221N330066N330066 n A L OOS . . . .OSS . . . .OSS . . . . a C P NNS1100NSS1111NSS1111
)
O L j N A E N B K OM 744527400000000000000 (
/ RP 3611983 L OP 075 46 m W B 1 u A n T i I RM m N EC u T . 066666696862620528006 l F S 33333337856356620197 a -
O UP 0666666 . . . . . . . . . . . . .
LE . . . . . 87679790998675 d 2 S CS 111111 1 e 2 T, ~t L a -
U H 666666622222224444444 r S C . 333333333333335555555 o E TM 66666660000000 . . . . . . . b R IC . . . . . . . . . . . . . 2222222 P 11111112222222 r o
f
.% 666666655555559999999 R 444444433333332222222 d NT . . . . . . . . . . . . . . . . . . . . . e EW 222222222222224444444 t o
n T
N IIIIVIX 56823892389R12 s E IIXIIII 22333331 011 i M I IXVX 1 I X X B R 11111112222222 E #######f#f#f## %
P WWWWWWWLLLLLLLLLLLLLL X &&&&&&&NNNNNNNNNNNNNN .
E BBBBBBBPPPPPPPPPPPPPP t w
' E e S . 123456789012345678901 h AO 111111111122 T CN *
!l
REFERENCEU TO APPENDIX A
- 1. " SCALE: A Modular Code System for Performing Standardized Computer Analysis for Licensing Evaluation, Vols 1-3,"
NUREG/CR-0200 (Rev'. 2), U.S. Nuclear Regulatory Commission (June 1983).
Distributed by the Radiation Shielding Information Center, P.O. Box X, Oak Ridge, TN 37830.
- 2. " Nuclear Criticality Safety in Operations with Fissionable Materials at Fuels and Materials Facilities," USNRC Regulatory Guide 3.4, Rev. 2 (March 1986).
- 3. M. N. Baldwin et al., " Critical Experiments Supporting Close Proximity Water Storage Of Power Reactor Fuel," BAW-1487-7, The Babcock & Wilcox Company, (July 1979).
- 4. S. R. Bierman, E. D. Clayton and B. M. Durst, " Critical Separation Between Subcritical Clusters of 2.35 Wt. % 235U Enriched'UO2 Rods in Water with Fixed Neutron Poisons," j PNL-2438, Battelle Pacific Northwest Laboratories, (1977).
- 5. S. R. Bierman, E. D. Clayton and B. M. Durst, " Critical Separation Between Subcritical Clusters of 4.29 Wt. % 235U Enriched UO2 Rods in Water with Fixed Neutron Poisons,"
NUREG/CR-0073 (1978).
- 6. W. Marshall, P. D. Clemson and G. Weber, " Criticality Safety Criteria," ANS Transactions Vol. 35, pp. 278-279 (1980).
- 7. D. G. Napolitano, F. L. Carpenito and P. J. Rashid,
" Validation of YAEC Criticality Safety Methodology," ANS Transactions Vol. 56, pp. 325-327 (1988).
L_-______--. ._
Attachment 2 to AECM-89/0037 GGNS-1 HDSFSR BORAFLEX SURVEILLANCE' PROGRAM In order to verify that the GGNS-1 High Density Spent Fuel Storage Racks' (HDSFSR) neutron absorbing material (Boraflex)) remains within the analyzed parameters, a Boraflex surveillance program will be established. This progrgm will use neutron Blackness Testing to determine the size and frequency distributions of gaps in the Boraflex panels. The results of these measurements will be compared _to the assumptions used in the then current criticality safety analysis to verify that the 0.95 k-effective criticality safety limit is not compromised.
Blackness testing is typically performed using a specially designed Blackness Test instrument containing a fast neutron source and thermal neutron detectors.
The test instrument is lowered into each rack cell to be tested and slowly moved axially to scan the entire cell. Fast neutrons from the source pass through the test instrument walls and are thermalized in the water of adjacent rack cells. If Boraflex is not present or is significantly degraded, reflected neutrons will pass back through the rack wall and register as an increase in the count rate of the detectors inside the test instrument. If the Boraflex panel is intact, the reflected neutrons will be absorbed and no change in the count rate will occur.
This methodology is intended to be employed at GGNS Unit 1. All surveillance will be conducted in the same area of the spent fuel pool. This area is an 8 x 13 array of spent fuel pool storage locations beginning at cell HH-20 and ending at V-27. It is the site of the Boraflex gap measurements performed during Cycle 3. Surveillance will be conducted once per cycle with the first taking place in Cycle 4. During the refueling outage prior to each surveillance, freshly discharged spent fuel will be loaded into' the surveillance area. Following 10 to 14 months of irradiation the fuel will be removed and at least 50 rack cells will be measured using a neutron Blackness Test. An evaluation under 10CFR50.59(a)(1) will be performed in support of this test program.
Loading freshly discharged fuel into the: surveillance area each cycle will induce a significantly higher gamma fluence in the surveillance area than in other areas of the rack. Since gamma fluence has been strongly correlated to i gaps in Boraflex (1), the surveillance area will lead other areas of the racks l in the formation of gaps.
1 Since the size and frequency of Boraflex gaps are projected to reach equilibrium values during the Cycle 4 irradiation period, an assessment of the l long term surveillance requirements will be performed following the Cycle 5 surveillance. This assessment will be based upon both GGN$-1 specific measurements and other industry data. It is expected tMc further measurements will not be required once Boraflex behavior is fully characterized in a spent fuel environment. Current equilibrium gap behavior is projected based upon a harsher in-reactor environment (1). This projection is expected to bound the GGNS-1 Boraflex behavior. If these measurements indicate the need for a surveillance program beyond Cycle 5, then the surveillance area may be loaded with previously discharged fuel in order to prevent the development of unrealistically high gamma fluence levels.
J14 MISC 89021701 - 1
l Attachment 2
, to AECM-89/0037 In summary, this surveillance program will utilize a test area which leads other areas of the rack in the formation and growth of gaps in Boraflex panels.
The results of this program are expected to verify that the criticality analysis assumptions are bounding and to provide a more realistic characterization of Boraflex gap behavior in an in-service rack environment.
It is projected that two surveillance, with the first in Cycle 4, will demonstrate that the formation and growth of gaps have reached equilibrium and that further surveillance will not be necessary.
Raference:
- 1) "An Assessment of Boraflex Performance in Spent-Nuclear-Fuel Storage Racks", EPRI Report NP-6159, December 1988.
J14 MISC 89021701 - 2
_ _ _ _ _ - _ - _ _ _ _ _ - - _ _ _ _ _ _ _ _ _