ML081820352

Report 0000-0035-7327-SFP, Rev. 2, GNF2 Spent Fuel Storage Rack Criticality Analysis for Peach Bottom Atomic Power Station, Units 2 & 3.

ML081820352
Person / Time
Site:	Peach Bottom
Issue date:	06/30/2008
From:	Global Nuclear Fuel - Americas
To:	Office of Nuclear Reactor Regulation
References
0000-0035-7327-SFP, Rev 2
Download: ML081820352 (46)

Text

ENCLOSURE 6 GNF2 SPENT FUEL STORAGE RACK CRITICALITY ANALSIS FOR PEACH BOTTOM ATOMIC POWER STATION, UNITS 2 & 3 Non-Proprietary Version

Global Nuclear Fuel A Joint Venture of GE, Toshiba, & Hitachi 0000-0035-7327-SFP Revision 2 Class I June 2008 Non-Proprietary Version GNF2 Spent Fuel Storage Rack Criticality Analysis for Peach Bottom Atomic Power Station Units 2 & 3

0000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion IMPORTANT NOTICE REGARDING CONTENTS OF THIS REPORT Please Read Carefully The only undertakings of Global Nuclear Fuel (GNF) respecting information in this document are contained in the contract between Exelon and GNF, Nuclear Fuel Fabrication and Related Components and Services Contract, as amended to the date of transmittal of this document, and nothing contained in this document shall be construed as changing the contract. The use of this information by anyone other than Exelon, or for any purpose other than that for which it is intended is not authorized: and with respect to any unauthorized use, GNF makes no representation or warranty, express or implied, and assumes no liability as to the completeness, accuracy or usefulness of the information contained in this document, or that its use may not infringe privately owned rights.

NON-PROPRIETARY NOTICE This is a non-proprietary version of the document, from which the proprietary information has been removed. Portions of the document that have been removed are identified by white space within double square brackets, as shown here (( )).

ii

0000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion Contents

1. INTRODUCTION AND

SUMMARY

1-3

2. METHOD OF ANALYSES 2-1 2.1 Cross Sections 2-1 2.2 Geometry Treatment 2-1 2.3 Validation and Computational Basis 2-1

3. ANALYSIS & ACCEPTANCE CRITERIA 3-1 3.1 Criticality Analysis 3-1 3.2 Conformance with USNRC Regulatory Guide 1.13 and ANSI/ANS-57.2-1983 3-2 3.3 Definitions 3-3 3.4 Treatment of Biases and Uncertainties 3-4

4. FUEL DESIGN BASIS 4-1 4.1 GNF2 Fuel Lattice Composition 4-1 4.1.1 Additional GNF2 Fuel Lattice Composition 4-1 4.2 Effect of Vanished/Dominant Lattices on In-rack Neutron Multiplication 4-1 4.2.1 Additional Effects of Neutron Multiplication Factor on Dominant Lattice 4-2 4.3 Infinite Lattice Benchmark Uncertainty 4-2

5. SPENT FUEL STORAGE 5-1 5.1 Description of Spent Fuel Storage Rack 5-1 5.2 Normal Configuration 5-2 5.2.1 Analytical Model 5-2 5.2.2 Results 5-3 5.3 Accident/Abnormal Configuration 5-7 5.3.1 Analytical Model 5-7 5.3.2 Results 5-7 5.4 Uncertainty Analyses 5-8 5.4.1 Analytical Model 5-8 5.4.2 Results 5-10 5.5 Boraflex Degradation 5-10

6.

SUMMARY

6-1

7. REFERENCES 7-1 Appendix A INFINITE LATTICE BENCHMARKS A-1 Appendix B IN-CORE K_ CRITERIA METHODOLOGY B-1 Appendix C CRITICALITY ANALYSIS FOR FUEL INSPECTION C-1 I11

0000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion Tables Table Title Page Table I Material Isotopic Compositions (atoms/bnocm) 4-5 Table 2 Comparison of Geometric and Material Data for GNF2 and GE14 Fuel Lattices 4-7 Table 3 Assumed Spent Fuel Storage Rack Dimensions 5-5 Table 4 In-core versus Nominal In-rack Eigenvalues for Evaluated Benchmark Lattices (20'C) 5-6 Table 5 PBAPS Criticality Safety Analyses Summary 5-13 Table 6 Peak Cold Uncontrolled Lattice Reactivity 6-2 Table A I - GNF2 Results A-3 iv

0000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion Figures Figure Title Page Figure 1 GNF2 Fuel Lattice Configuration 4-4 Figure 2 Example of an Infinite D-Lattice Simulation Model 4-6 Figure 3. PBAPS Boraflex Spent Fuel Storage Rack Cell 5-1 Figure 4 Spent Fuel Storage Rack Cell " 5-4 Figure 5 Spent Fuel Storage Rack, Eccentric Loading 5-9 Figure 6 NETco. Reactivity Comparison and Boraflex Degradation Penalty 5-11 Figure A 1. PBAPS Boraflex Rack In-Core vs. In-Rack Eigenvalues A-2 V

O000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion Revision Status Revision Section Description of Change Number 0, 1 Internal Documents 2 Initial USNRC Submittal I-1

0000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion Executive Summary This report describes the criticality analyses and results for the Peach Bottom Atomic Power Station (PBAPS) spent fuel storage pool for lOx10 GNF2 fuel lattice bundles in the Westinghouse Boraflex (Reference 1) spent fuel storage racks. This analysis was performed with the Monte Carlo neutron transport program MCNP0IA (Reference 2). Fuel bundle geometries were assumed with uniform lattice enrichments up to 4.90 wt% 235U using both full and part length fuel rods and including representative placement and numbers of Gadolinia rods with a range of loadings (weights). The spent fuel storage rack-analyzed lattices were evaluated at their respective cold, uncontrolled, peak exposure-dependent reactivity statepoints at both zero and 70 percent void histories. Current design basis was set to specifications concluding a 70% void history. A 4.90 wt % 231U lattice and 2 wt% gadolinium concentrations in 13 specified rods with initial uniformly distributed fissile inventories and burnable absorber, at the peak cold, exposure-dependent eigenvalue statepoint (with associated exposure-dependent isotopics) is the limiting lattice design to satisfy the spent fuel storage rack eigenvalue criteria, in conjunction with the design parameters and calculational assumptions applied. The associated in-core peak eigenvalue (k-) of 1.318 is a TGBLA06A production lattice physics computed value, with benchmark uncertainties included, and evaluated in conjunction with the 4.90 wt% 235U lattice average enrichment establishes the spent fuel storage rack design limit for the PBAPS. The analysis conditions and assumptions used are in compliance with the requirements contained in the USNRC Regulatory Guide 1.13, Spent Fuel Storage Facility Design Basis, Rev. 2 (Reference 3), and ANSI/ANS-57.2-1983, Design Requirements for Light Water Reactor Spent Fuel Storage Facilities at Nuclear Power Plants (Reference 4).

The PBAPS spent fuel storage racks were analyzed with GNF2 fuel and shown to have a keff less than 0.95 for both normal and credible abnormal configurations. The normal configurations included displacement of stored fuel assemblies within the ranges permitted by the rack manufacturing tolerances at a conservative pool water density corresponding to a temperature of 20'C. The abnormal configurations considered the storage of spent fuel without flow channels and with off-center bundle locations. The PBAPS Westinghouse manufactured Boraflex spent fuel storage racks satisfy the reactivity requirements for all storage conditions with GNF2 fuel having an associated in-core peak eigenvalue (kL) of no greater than 1.318 inserted. Additionally, a delta (ker) penalty value for Boraflex degradation was implemented on all simulated in-rack exposures (Reference 18). The conclusions of this analysis are valid only for the assumptions and information provided to GNF by PBAPS for this purpose. Should it be determined that any of the assumptions regarding storage rack design, construction, manufacturing tolerances or material compositions be different than those assumed in this document, this analysis must be re-evaluated to assess any possible impact on the final results that these changes might produce.

This GNF2 analysis bounds all previously manufactured GNF lattice designs for Peach Bottom Units 2 and 3.

1-2

O000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion

1. INTRODUCTION AND

SUMMARY

The PBAPS high-density Boraflex fuel storage rack system for spent fuel has been analyzed with GNF2 fuel bundles in order to demonstrate that system's ability to satisfy its design basis licensing requirements while loaded with these fuel products. The analyses were performed using the MCNPOIA Monte Carlo neutron transport computer program, which has been qualified for this application (Reference 2). The fuel lattice geometries used are consistent with typical l0x10 fuel designs with a uniform enrichment of 4.90 wt% 235U and varying gadolinium concentrations, consistent with the plant's core geometry (BWR-4). The spent fuel storage rack simulations incorporated lattices with depleted burnable poison and explicit fission product inventories associated with lattice peak cold, exposure-dependent in-core statepoints, and desired void histories for all designs considered. The fuel assemblies considered contained both full and part length fuel rods. Normal and abnormal spent fuel storage rack configurations were evaluated with two-dimensional geometry models. The analysis conditions and assumptions used are in compliance with the requirements contained in the USNRC Regulatory Guide 1.13, Spent Fuel Storage Facility Design Basis, Rev. 2 (Reference 3), and ANSI/ANS-57.2-1983, Design Requirements for Light Water Reactor Spent Fuel Storage Facilities at Nuclear Power Plants (Reference 4).

The PBAPS Boraflex storage racks maintain a subcritical multiplication factor, kff, for all normal and abnormal material/geometry scenarios for GNF2 fuel bundles and for all GNF/GE fuel bundle products which have been utilized in the Peach Bottom Units 2 and 3. In order to insure that specific bundles are acceptable for storage in the Boraflex spent fuel storage racks, the following three items should be considered:

1. Identify the unique lattices in each assembly (bundle) design.

2. Determine the maximum cold 68F (20C) uncontrolled k-. results from standard TGBLA06A lattice physics calculations for each unique lattice in the assembly.

3. Ensure that the k-. values obtained from Step 2 for each lattice are less than the rack-specific in-core peak eigenvalue (kL,) limit of 1.318 at 70 percent void histories. Such fuel can be loaded in any configuration in the storage racks.

Note that lattice physics codes other than TGBLA06A may be used to perform the evaluation, subject to a determination of an appropriate kL. limit having been determined for the particular lattice physics code to be used.

1-3

0000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion

2. METHOD OF ANALYSES The documented fuel storage criticality calculations were performed using the MCNP0IA (Reference 2) computer code. MCNPOIA is a Monte Carlo program for solving the linear neutron transport equation as a fixed source or an eigenvalue problem in three space dimensions. It implements the Monte Carlo process for neutron, photon or electron transport, or coupled transport involving all these particles, and can compute the eigenvalue for neutron-multiplying systems. For the present application only neutron transport was considered.

2.1 Cross Sections MCNPOIA uses point-wise (i.e., continuous energy) cross section data, and all reactions in a given cross section evaluation (e.g., ENDF/B-V) are considered. For the present work, thermal neutron scattering with hydrogen was described using an S(a,p) light water thermal scattering kernel. The cross section tables include all details of the ENDF representations for neutron data. The code requires that all the cross sections be given on a single uniform energy grid suitable for linear interpolation; however, the cross section energy grid varies from isotope to isotope. The libraries include very little data thinning and utilize maximum resonance integral reconstruction error tolerances of 0.001 %.

2.2 Geometry Treatment MCNPO IA implements a robust geometry representation that can correctly model complex components in three dimensions. An arbitrary three-dimensional configuration is treated as geometric cells bounded by first and second-degree surfaces. The cells are described in a Cartesian coordinate system and are defined by the intersections, unions and complements of the regions bounded by the surfaces. Surfaces are defined by supplying coefficients to the analytic surface equations or, for certain types of surfaces, known points on the surfaces. Rather than combining several pre-defined geometrical bodies in a combinatorial geometry scheme, MCNPOIA has the flexibility of defining geometrical shapes from all the first and second-degree surfaces of analytical geometry and then combining them with Boolean operators. The code performs extensive checking for geometry errors and provides a plotting feature for examining the geometry and material assignments.

2.3 Validation and Computational Basis MCNPOIA has been compared to critical experiments for different temperatures and moderator-to-fuel ratios and many of these experiments were consistent with BWR geometries and materials. The critical experiments to which MCNPOIA has been compared are:

2-1

O000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion Cross Section Evaluation Working Group (CSEWG) thermal reactor benchmark problems:, PNL-1 and PNL-2; Small Core Criticals with and without Poison Curtains; Small Core Criticals with and without Burnable Absorbers LE Er 1]

2-2

0000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion

3. ANALYSIS & ACCEPTANCE CRITERIA 3.1 Criticality Analysis The design of the spent fuel storage racks provides for a subcritical multiplication factor for both normal and abnormal storage conditions. For all conditions, the storage rack eigenvalue must be _<0.95. Normal conditions exist when the fuel storage racks are located in the pool and are covered with a normal depth of water (about 25 ft above the stored fuel) for radiation shielding and with the maximum number of fuel assemblies or bundles in their design storage positions. The spent fuel is covered with water at all times by a minimum depth required to provide sufficient shielding. An abnormal condition may result from accidental dropping of equipment or damage caused by the horizontal movement of fuel handling equipment without first disengaging the fuel from the hoisting equipment. To meet the requirements of General Design Criterion 61 (Reference 9), geometrically-safe configurations of fuel, stored in the spent fuel array, are employed to assure that the eigenvalue will not exceed 0.95 under all normal and abnormal storage conditions. The geometry of the spent fuel storage array must additionally ensure that the eigenvalue will be _<0.95 under optimum moderation conditions.

To ensure that the design criteria are met, the following normal and abnormal spent fuel storage conditions were analytically considered:

" Normal geometry configuration in the spent fuel storage array

" Temperature and water density effects

• Abnormal locations of a fuel assembly, including non-channeled bundle storage

" Eccentric fuel assembly positioning

" Dropped fuel assembly

" Fuel rack lateral movement No limitation was placed on the size of the spent fuel storage array from a criticality standpoint, since all calculations are performed on an infinite basis.

To assure the true reactivity would always be less than the calculated reactivity, the following conservative assumptions are made:

3-1

0000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion

'((i I]

))

))

.. ((I I]

3.2 Conformance with USNRC Regulatory Guide 1.13 and ANSI/ANS-57.2-1983 The basic regulatory requirements for assuring margin to criticality for spent fuel storage at PBAPS are defined in the USNRC Regulatory Guide 1.13 (Reference 3), with design basis and criticality specifications defined in ANSI/ANS-57.2-1983 (Reference 4).

The margin requirement is keff _<0.95 for normal and abnormal operation with tolerances and computational uncertainties taken into account. Such Nuclear Criticality Analyses are subject to the following restrictions (ANSI/ANS-57.2-1983 [6.4.2.2]):

" The most reactive fuel assembly anticipated for insertion into the storage must be assumed;

• For fuel assemblies that incorporate a distribution of discrete fissile inventories within the assembly, the conservatively determined mean fissile inventory shall be assumed;

• Credit for fixed neutron poisons (such as integral burnable absorbers - Gadolinia) may be taken;

• The analysis must consider the location of assemblies in transit to the storage racks or in the fuel inspection stand;

• Neutron poisons incorporated in the storage structure may be credited in increasing the margin to criticality.

3-2

0000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion The scope of the Nuclear Criticality Analysis shall consider credible abnormal occurrences, as well as normal conditions (ANSI/ANS-57.2-1983 [6.4.2.1]).

The margin requirement is determined in part by establishing a Akuncertainty contribution to the storage rack eigenvalue, as detailed in Section 3.4 of this report.

3.3 Definitions The following terms are used in this report:

• Fuel Assembly - is a complete fuel unit consisting of a basic 1OxlO fuel rod structure, including large central water rods. Several shorter rods may be included in the assembly. These are called "part length" rods. A fuel assembly includes the fuel channel.

" Gadolinia - The compound Gd 2 0 3 . The gadolinium content in integral burnable absorber fuel rods is usually expressed in weight percentage of Gadolinia.

" Integral Burnable Absorbers - are neutron poisons used to reduce the reactivity of a highly enriched lattice, that are blended with fissile material (i.e., U0 2) in the fuel pellets. In a typical BWR, the absorber of choice is Gadolinia.

" Lattice - An axial zone of a fuel assembly within which the nuclear characteristics of the individual rods are unchanged.

" Dominant Lattice - An axial zone of a fuel assembly typically located in the bottom half of the bundle above the power shaping zone region.

" Vanished Lattice - An axial zone of a fuel assembly typically located in the upper half of the bundle above the dominant zone region.

" Sub-critical - having a neutron multiplication factor (ker or k-) less than 1.0 after taking into account statistical uncertainties and biases. In criticality safety analyses in which Monte Carlo codes are used, subcriticality is usually demonstrated by showing that the maximum keff is less than some offset from unity - for example, is less than 0.95. Biases are considered to be negative if critical benchmarks are under-predicted (i.e., result in calculated multiplications less than unity).

3-3.

O000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion 3.4 Treatment of Biases and Uncertainties The eigenvalues reported herein contain an assessment for computer code bias and uncertainty that is consistent with Reference 4 and formalized as follows:

(1) kr.x = km. + Akbias + Akuncertainty In this equation, the terms are defined as:

kmc - eigenvalue from Monte Carlo calculation; Akbias - critical benchmark bias for Monte Carlo code (MCNPO 1A versus critical experiments); the bias due to Boraflex degradation, and AkU2 = Ak associated with non-channeled analysis Akuncerlainty -

Akul = Problem specific relative error (17)

Aku 3 = Ak associated with eccentric assembly location (diagonal off-set)

AkU4 = Ak associated with uncertainty in fuel enrichment AkU5 = Ak associated with uncertainty of lattice spacing Aku 6 = Ak associated with uncertainty in SS box thickness Aku 7 = Ak associated with bias from Boraflex degradation Aku 8 = Ak associated with decreased pool water temperature (4°C)

The standard deviation of the Monte Carlo eigenvalue simulations (Akul) is the la eigenvalue uncertainty directly associated with the individual storage rack lattice-calculated k_(s). This standard deviation is computed by MCNPO1A relative to the mean eigenvalue, tallied for a large number of neutron batches, or cycles, and is part of the normal MCNPOIA output associated with an eigenvalue solution. The 1a value incorporated for this uncertainty analysis is extracted from the nominal GNF2 design basis lattice evaluation output. An ambient pool water temperature of 20'C was assumed.

The non-channeled analysis (Aku 2 ) simulates the storage of the most reactive lattice in an infinite storage rack array, without flow channels, and with nominal specifications of the associated storage rack material/geometry parameters. This geometry is treated as a bias if it results in a positive reactivity impact.

3-4

0000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion The eccentric assembly location analysis (AkU3) simulates the storage of the most reactive lattice, including flow channels, offset to an extreme inner comer (diagonal off-set) of an associated storage rack cell, in an infinite storage rack array, and with nominal specifications of the associated storage rack material/geometry parameters.

The uncertainty in the fuel enrichment analysis (Aku 4) simulates the storage of the most reactive lattice, including flow channels, with nominal specifications of the associated storage rack material/geometry parameters, ((

The uncertainty in lattice spacing analysis (Aku5 ) simulates the storage of the most reactive lattice, including flow channels, with nominal specifications of the associated storage rack material/geometry parameters, with a 1 mm decrease in the nominally specified storage cell center-to-center spacing of 15.9512 cm. This value of 1 mm is consistent with the GEl4 analysis for PBAPS. Decreasing the lattice spacing, or pitch, in an over-moderated system will increase the system reactivity, and the simulation was performed to quantify the expected result.

The uncertainty in the stainless steel box thickness analysis (Aku 6) simulates the storage of the most reactive lattice, including flow channels, with nominal specifications of the associated storage rack material/geometry parameters, with a 10% (0.0075") decrease in the associated nominally specified inner steel cell wall thickness of 0.075". The 10% reduction is consistent with the previous analysis for GE 14.

To account for the reactivity effects of Boraflex degradation, a reactivity penalty will be applied in the form of a bias (Aku7). This bias is based on previous analyses (Reference 10) performed for Exelon by another vendor. Confirmatory calculations were performed for the bounding GNF2 lattice to verify that the reactivity bias obtained from Reference 10 could be conservatively applied to GNF2.

The uncertainty in the storage rack water (moderator) temperature and density analysis (Aku 8) simulates the storage of the most reactive lattice, including flow channels, with nominal specifications of the associated storage rack material/geometry parameters, with a 4°C ambient water temperature, and an associated density of 62.426 lbs./ft3, providing a bounding water-temperature sensitivity evaluation.

3-5

0000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion

4. FUEL DESIGN BASIS 4.1 GNF2 Fuel Lattice Composition Criticality safety analyses to determine stored fuel reactivity were performed using the GNF2 1Oxl0 fuel design for the PBAPS spent fuel storage racks. The GNF2 fuel lattice configuration is a lOxl0 fuel rod array minus eight fuel rods that have been replaced with two large water rods as shown in Figure 1. The limiting evaluated lattice(s) incorporated multiple Gadolinia rods (1.0 wt% to 4.0 wt%) with initial uniform enrichments of up to 4.90 wt% 235U (fully rodded and vanished lattice). The Gadolinia loadings (number of rods and weights) utilized have been determined consistent with expected bundle designs.

The particular lattice design which could be demonstrated safe in the storage racks (including all uncertainties) had in-core k- =1.318 at the corresponding 70 percent void history. The associated material isotopic compositions used in the analysis are shown in Table 1. The GNF2 bundle dimensions are indicated in Figure 1 and Table 2. Also given in Table 2 is a comparison to the standard GE14 design for comparison. Conservation of mass was used to smear the U0 2 pellet isotopics over the pellet-to-cladding gap. The total U0 2 inventory is consistent with the design pellet stack density of [1 1]

theoretical density for GNF2 fuel. The GNF2 interactive channel geometry was maintained. An example of a GNF2 lattice is depicted in Figure 2 showing the geometry configuration with the Gadolinia rods readily identifiable by their respective multi-ring definitions (also see Appendix A). Two-dimensional nuclear simulations were performed for a number of GNF2 lattice models, using the GEH lattice physics codes, TGBLA06A (Reference -12). Configurations chosen for subsequent analysis and inclusion in this report yielded associated in-core eigenvalues (k-), as a function of void, temperature and exposure, utilizing the associated TGBLA06A resolved lattice statepoints.

4.1.1 Additional GNF2 Fuel Lattice Composition Recent criticality calculations on the GNF2 lOxl0 fuel designs incorporated the lattices requirements specified by NETco (Reference 18). The gadolinium rod layout was consistent with Figure 5 at a dominant lattice zone specification. The lattice consisted of uniform enrichments of 4.9 wt% 135Xe and 13 rods containing 2.0 wt% Gadolinium concentration. Two-dimensional lattice physics code, TGBLA06A (Reference 12) established appropriate in-core eigenvalues (k,), as a function of void fraction, temperature and exposure.

4.2 Effect of Vanished/Dominant Lattices on In-rack Neutron Multiplication For this analysis, TGBLA06A exposure-dependent in-core eigenvalues (ko,) for both dominant and vanished lattices were computed and associated MCNPOIA eigenvalues (Lk) were produced at each lattice's respective peak reactivity exposure point. This calculational suite included the GNF2 design basis lattices in Lead Use Assembly (LUA) bundle (( )) (dominant - ((

)), and vanished - []). The TGBLA06A peak, cold in-core eigenvalues (kL) were 1.1596 (dominant) and 1.1625 (vanished) at 17 and 15 GWd/st exposure respectively. The associated nominal in-rack eigenvalues (k-) for these lattice statepoints were 0.7843 (dominant) and 0.7827 (vanished), respectively (excluding calculational bias). This result 4-1

O000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion indicates that the dominant lattice's slightly lower H/U ratio (relative to the vanished lattice) produces a slightly higher in-rack eigenvalue than the dominant lattice, with a lower in-core eigenvalue.

A fuel product-dependent rack efficiency expression, defined as the ratio of a particular lattice statepoint in-rack eigenvalue (k-o) divided by the associated lattice nominal in-core eigenvalue (k-), allows for a straightforward comparison of both a storage rack's criticality response to different fuel products and to varying lattice designs within a particular fuel product. By this definition, a storage rack's capability to suppress reactivity for a specific lattice design can be readily compared to other fuel products, or lattice designs within the same fuel product line. A lower rack efficiency ratio implies an increased reactivity suppression capability (for a given product/lattice design), relative to an alternate design with a higher rack efficiency ratio. In other words, when comparing two racked statepoints, the lower the efficiency ratio, the more capable (better) the system. By comparing the rack efficiency ratios for the two nominal 10xl0 GNF2 lattice designs evaluated, the sensitivity of the in-rack reactivity to lattice design may be gleaned.

A comparison of ratios for the indicated design basis lattices (0.676 (dominant); 0.673 (vanished))

indicates that the there is very little difference in rack efficiency ratios between the dominant and vanished lattices analyzed in this study. This is also seen by examination of Figure A I which shows the dominant and vanished lattice reactivity curves essentially on top of each other over a wide range of in-core eigenvalues. As a result, the lattice with the highest in-core eigenvalue (a variant of dominant lattice 6939) will form the GNF2 licensing design basis for this spent fuel storage rack system. Note that previously the most reactive lattice configuration evaluated in the storage rack cell simulations (dominant) yields an in-core peak eigenvalue (k-) of 1.359 at zero void and 4.9 wt% U-235 and I wt%

Gadolinium concentration, without benchmark uncertainties included. Specifically, the (k-) values are directly reported from the TGBLA06A cold, uncontrolled summary edits. While single point evaluations could produce misleading results (rack efficiency ratios) for a specific lattice configuration, given the coarse fidelity of the discrete exposure increments employed, multiple lattices at various exposure points have been herein evaluated to characterize the vanished and dominant lattice efficiency ratios.

4.2.1 Additional Effects of Neutron Multiplication Factor on Dominant Lattice Dominant lattice exposure eigenvalues were calculated by the TGBLA06A lattice physics engine and verified by MCNPOIA lattice core simulations. Zero and 70 percent void histories were compared with the new bundle specifications. The simulated TGBLA06A peak, cold in-core zero and 70 percent void eigenvalues (k-) were 1.345 (dominant) and 1.318 (dominant) at 8 GWd/st and 10 GWd/st exposures respectively. The associated nominal MCNPO1A in-rack eigenvalue (k1) for these lattice statepoints at zero and 70 percent voids were 0.916 (dominant) and 0.901 (dominant) correspondingly (excluding calculation bias).

4.3 Infinite Lattice Benchmark Uncertainty Nominally determined TGBLA06A infinite lattice eigenvalues (k) do not include uncertainties versus benchmark analyses. A (( )) has been determined for TGBLA06A cold eigenvalue solutions, relative to MCNPO1A benchmark lattice evaluations. The infinite lattice cold 4-2

0000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion eigenvalue uncertainty is an integral component of the maximum storage rack eigenvalue assessment, and accordingly the (( Ak- is included in the storage rack uncertainty roll-up of k ax.

For subsequent benchmarking purposes, several GNF2 lattice designs are evaluated, with each infinite lattice's k-. compared with an associated nominal storage rack k-, consistent with the in-core k- criteria methodology employed for this analysis and report, and those lattices are fully specified, along with the associated results, in Appendix A of this report.

4-3

0000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion

[r 1]

Figure 1 GNF2 Fuel Lattice Configuration 4-4

0000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion Table 1 Material Isotopic Compositions (atoms/bn-cm)

I IS TIC ;I.0 t%25 Moderator at 200 C Oxygen 3.3370E-2 Hydrogen 6.6734E-2 Cladding and Channel Zirconium 4.32392E-2 304L Stainless Steel*

Silicon 1.6728E-3 Manganese 1.4906E-2 Iron 5.6437E-2 Nickel 8.4036E-3 Manganese 1.71OOE-3 235 U0 2 at 4.90w% U (96.5% TD) 234u 1.0343E-5 235 U 1.1 165E-3 238U 2.1386E-2 Oxygen 4.5025E-2

• - Standard material composition taken from Reference 13.

4-5

0000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion Figure 2 Example of an Infinite D-Lattice Simulation Model O000000000 0'00() 0000 000000-000 OOO 1CO@O 010000001001 OOO00O000000

• ©0000000C 4-6

0000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion Table 2 Comparison of Geometric and Material Data for GNF2 and GE14 Fuel Lattices Parameter GNF2 GE14 Number of U0 2 Rods ((

Total Number of Gd-U0 2 rods Number of Water Rods Fuel Pin Pitch (cm)

Fuel Rod OD (cm)

Fuel Pellet OD (cm)

Cladding Thickness (cm)

Cladding Material U0 2 Pellet Stack Density (g/cc)

Fuel Bundle Channel Thickness (cm)

Channel Material Active Fuel Length (cm) ))

4-7

O000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion

5. SPENT FUEL STORAGE 5.1 Description of Spent Fuel Storage Rack The PBAPS design basis Boraflex storage rack cell assembly manufactured by Westinghouse (Reference 1) consists of a large 304 stainless steel structure composed of a series of square vertical tubes (cells). These tubes contain 0.08 1" thick Boraflex panels sandwiched between a 0.075" SS inner cell wall and a 0.020" SS outer wrapper. The Boraflex containing cells are arranged in a checkerboard pattern with the space between a 4-cell group forming a fifth bundle storage location with a center-to-center cell pitch of 6.280 inches.

Rack array sizes ranging from 9x 14 up to 19x20 are placed adjacent to one another in the spent fuel pools of both PBAPS Units 2 and 3. The rack employs thermal neutron absorption in the B- 10 of the Boraflex as the primary mechanism of reactivity control. A schematic of a single storage rack unit-cell is shown in Figure

3. All information pertaining to the geometry and material composition of the racks was taken from the original design basis document (Reference 1) for these racks given by the supplier Westinghouse.

6.280' UNIT CCLL OF INFINITE ARR* Y 6.280" PERE

-- 4.900'kAEX-ý 6.070" 0.085" U

UETAIL 'A' S6.120" 0.090" Ga U 0,075" Inner Cell Wall 0 Boraflex L 0.020" Wrapper Figure 3. PBAPS Boraflex Spent Fuel Storage Rack Cell 5-1

0000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion 5.2 Normal Configuration The normal storage environment includes bundles with flow channels, within 304L stainless steel sleeves, in a normally pitched spacing arrangement, with a conservative ambient water temperature of 20'C. In the associated analysis simulations, the fuel material inventories include pin-by-pin isotopic representations of the infinite lattice loading at the peak cold, exposure-dependent reactivity statepoint.

For the two dimensional lattices evaluated, with pins initially loaded with uniform enrichments of 235U and fixed initial loadings of Gadolinia rods, the peak cold, exposure-dependent reactivity statepoints are considered (burned under hot, zero-void, full power conditions). Consideration of an arrangement of in-rack spent fuel assemblies containing burnable absorbers and fission products (versus no poisons) is necessary to demonstrate subcriticality compliance for highly enriched fuel assemblies in HDFS systems, with regard to USNRC Regulatory Guide 1.13 (Reference 3) and ANSI/ANS-57.2-1983 (Reference 4).

5.2.1 Analytical Model Due to the design of the racks, a 2-D, axially infinite checkerboard lattice was created. The first lattice cell contained the center storage rack cell position (seen in Figure 3) composed of the inner stainless steel liner and part of the Boralfex panel region. The second (adjacent) storage cell position contained the remaining Boraflex region and the outer stainless steel structure. The design of the checkerboard rack necessitated this type of modeling since the periodic line of symmetry for the square lattice (6.280" pitch) occurred in between the two adjacent cells. A 23x23 checkerboard array of alternating cells was created and periodic reflecting boundary conditions were imposed on the four peripheral X and Y boundary planes to accurately simulate an infinite checkerboard storage pattern. Axially reflective (spectral) boundary conditions are imposed in the Z plane, conservatively simulating an axially infinite pool geometry.

Previous studies have based spent fuel storage rack analytical calculations on beginning-of-life material configuration models without incorporating burnable absorbers. It was recognized early in the process of performing this study, that for an infinite array of highly enriched GNF2 bundles in the PBAPS Boraflex spent fuel storage racks, the pertinent reactivity criteria could not be satisfied without incorporating burnable absorbers in some of the fuel pins.

As a result, a number of conservative material configuration scenarios, incorporating burnable absorbers, have been herein defined for this spent fuel storage rack analytical study. The conservative interpretation stems from the selection of peak, exposed reactivity statepoints, ignoring significant poisons in the form of "lumped" fission products, and ignoring any 135Xe inventory which is assumed to have decayed away.

Pin-by-pin isotopic specifications were generated by burning the GEH lattice physics production code (TGBLA06A), with hot operating fuel and moderator temperatures, at the zero-void moderator condition, to peak reactivity anomaly statepoints. For conservatism, the lattice physics generated "lumped" fission product was neglected in the Monte Carlo simulations (non beginning-of-life configurations), adding an additional degree of conservatism to the spent fuel storage rack calculations.

5-2

O000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion The vast majority of the initially loaded Gadolinia has burned out by these exposure points, so the remaining Gadolinia inventory is small. The 135Xe inventory is assumed to have decayed away completely, and is ignored for all exposed spent fuel storage rack analyses.

To determine the loading arrangement of pins for the subject evaluation, lattice configurations (number and location of Gd rods) were chosen consistent with anticipated GNF2 bundle design for PBAPS GNF2 New Fuel Introduction, with regard to evaluating lattice peak cold, exposure-dependent eigenvalues against the anticipated storage rack reactivity capacity. The identified configurations were benchmarked with MCNPOIA calculations to determine the accuracy of the lattice physics predicted eigenvalues and to determine the acceptability of the lattice physics based methodology of establishing the respective peak reactivity points, without uncertainties. The Monte Carlo analyses confirmed the TGBLA06A's predicted statepoint design eigenvalues to within the documented 95%/95% tolerance uncertainty range of [I )), validating the design basis methodology and statepoint selection criteria for the pertinent material configurations. Additionally, the Monte Carlo results fall within the uncertainty ranges developed for the GNF2 fuel product line cold eigenvalue basis 95%/95% confidence level.

5.2.2 Results The normal configuration fully-loaded rack Monte Carlo simulation results indicate that the limiting lattice configuration (in-core kL. =1.321), in the nominally specified (material/geometry) spent fuel storage rack with 20'C, 70 percent void water, yields an in-rack eigenvalue below the (kmax) < 0.95 limit, with computational uncertainties taken into account, for an initial uniform lattice enrichment of 4.90 wt%

235 U with a 2 wt% gadolinium concentration.

The spent fuel storage rack limiting lattice, with a lattice enrichment of 4.90 wt% 235U and 2 wt%

gadolinium concentrations with the current GNF2 bundle design (see Figure 5). This yields a nominal MCNP0IA peak in-core eigenvalue (L) of 1.321 with an associated nominal peak in-rack eigenvalue (kin.rack) of 0.9018 (+/- 0.00036), without bias for 20'C ambient pool water temperature. (Note that the fuel pin isotopics used in the Monte Carlo in-rack lattice simulations are identical to those produced by the TGBLA06A lattice bum calculations. There was no re-distribution of fissile (or non-fissile) inventory; the as-burned spatial isotopic pin inventories were exactly preserved at the respective exposure point(s)).

5-3

0000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion J*

I

-i1III

• ýEi I 3--1 I 010101010101010100 010001oo010101010 00000000000 100000000001

• oboooooooo 00000.00000 oooo0000o00010 1001000V @000 00000 000C

.0.0000 0100 000 00000 000__000001 oooojoooooo 00.0 0.0000 0.00 00000 0000000000 010000,00000 0000000000 10.01000,0010100 m

'41 00000000 "A-u (00000-00000 0000000 00 5ooooo~ooodl 001000,00000 0000000000W 1001000 000o00o000 00000 00000000001 00000 ON.1000I 1010000No 00 0 1 000 000000 1000 00000 O0~000100O 00.00000.0100 00000000100 0o000000 0000000000o 100QO000100010Q 1O01O001010101

]J Y

iI

.J l1

{I1 1 1 1 1 1 1 1I I II WL1111111111 Ii Figure 4 Spent Fuel Storage Rack Cell 5-4

0000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion Table 3 Assumed Spent Fuel Storage Rack Dimensions Storage Cell Pitch 15.9512 cm Storage Cell Liner Dimension 15.4178 cm Intermediate Storage Location Inner Dimension 15.5448 cm 304L Stainless Steel Inner Cell Wall 0.1905 cm 304L Stainless Steel Outer Wrapper 0.0508 cm Gap Between Inner Cell Wall and Wrapper 0.2286 cm Boraflex Panel Thickness 0.20574 cm Water Gaps on Either Side Of Boraflex Panel 0.01143cm 5-5

0000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion Table 4 In-core versus Nominal In-rack Eigenvalues for Evaluated Benchmark Lattices (20'C)

Lattice Lattice In-core In-core In-rack ID+ Type TBGLA06 MCNPO1A MCNPO1A 6939c Dominant 6941 c Vanished 6939b Dominant 6941b Vanished 6943a Vanished 6939a Dominant 6941 a Vanished 6943 Vanished 6942 Vanished 6939 Dominant 6941 Vanished 6940 Dominant

• 543212 MXDom-0%vh

• 543212 MXDom-70%vh 1]

'Note: GNF2 lattices with varying enrichments, Gd loadings (number, placement, weight) and peak reactivity exposure points are shown. Lattices IDs are given as 4-digit integers. For lattices specified as XXXXa, XXXXb, XXXXc, etc., these are variations of the most reactive dominant (6939) or vanished (6943) lattice obtained by changing the Gd wt% content in the rods.

235

• Note: GNF2 lattices with 4.90 wt% U enrichments and 2.00 wt% gadolinium concentrations (see Figure 5). Peak zero and 70 percent void history reactivity exposure points are shown.

5-6

O000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion 5.3 Accident/Abnormal Configuration The spent fuel storage rack abnormal configurations involve the following postulated scenarios, or occurrences, and are evaluated with MCNPOIA in the nominal storage rack cell model, where needed:

" Temperature increase/decrease of pool water

" Lateral movement of a rack module

" Misplacement of a fuel assembly

" An assembly dropped on top of the rack

" Eccentric fuel assembly positioning

" Other (non-channeled assemblies if appropriate) 5.3.1 Analytical Model The nominally specified material/geometry parameters for the storage rack cell were utilized, in conjunction with the limiting dominant GNF2 lattice to form the basis for the abnormal scenarios.

Individual parameters were perturbed accordingly, to appropriately simulate the physical changes associated with a given abnormal scenario.

5.3.2 Results 5.3.2.1 Temperature Increase of Pool Water All calculations in this analysis were performed for material densities and cross-sections corresponding to 20'C. In order to account for pool water temperature decrease effects, a cold water simulation at 4'C (the maximum theoretical water density temperature of 62.426 lbs/ft3 ) was performed. This increase in water density resulted in a reactivity increase of +0.0017 Ak in overall rack reactivity relative to the nominal design basis temperature model at 20'C.

5.3.2.2 Lateral Movement of a Rack Module For a finite spent fuel storage rack assembly, any enhanced separation between individual rack modules, or cells, results in a decrease of the actual system eigenvalue. The infinite array, close-packed storage cell configuration, forming the basis for all the calculations performed for this evaluation, establishes effective limiting eigenvalues. Considering that the actual PBAPS storage racks have some small separation between them, and that the Monte Carlo simulations used herein model individual storage cells in infinite arrays, the resulting solutions implicitly provide conservative estimates of the actual spaced storage rack modules in the pool.

5-7

0000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion 5.3.2.3 Misplacement of a Fuel Assembly; Dropped Assembly The spent fuel storage rack criticality calculations assumed an infinite array of storage cells, loaded with the most reactive lattice analyzed, and did not incorporate any radial or axial leakage. As a result, a model simulating a misplaced assembly adjacent to any plane, and in any orientation, would only begin to approach the limiting reactivity value evaluated for the nominal two-dimensional configuration.

Specifically, because of the distance between the top of the upper tie plate bale handle and the actual top of active fuel (a distance of >12"), any assembly that drops on the top of the racks is physically isolated from the remainder of the fuel, and thus the dropping of an assembly does not change the multiplication factor for the entire storage system as a whole.

5.3.2.4 Eccentric Fuel Assembly Positioning A fuel assembly is normally centered within an individual storage cell, with appropriate bottom fittings and spacers to prevent lateral fuel assembly motion. An eccentric position scenario was analyzed whereby a group of four bundles were each positioned in the extreme storage cell comer (diagonal off-set) of their respective positions in the center of the 23x23 periodic reflected array of bundles stored in their nominally centered positions. This case produced a positive reactivity delta of +0.0012 Ak, relative to the normal storage cell configuration result. This configuration is depicted in Figure 5.

5.3.2.5 Non-channeled Assemblies A fuel assembly is typically stored in the spent fuel storage rack within a zirconium flow channel. Since the storage racks are over-moderated to start with, eliminating the channel and replacing it with water produces a negative reactivity delta of -0.0030 Ak in the spent fuel storage rack eigenvalue (kmax),

relative to the normal storage cell configuration result. This perturbation is considered a bias.

5.4 Uncertainty Analyses The spent fuel storage rack criticality margin requirement is determined in part by establishing a Akuncertainty contribution to the storage rack eigenvalue, as detailed in Section 3.4 of this report. One of the contributions to the eigenvalue (k,.) uncertainty component stem from manufacturing tolerances associated with the fuel pellet enrichment. The Monte Carlo calculation relative errors contribute to the uncertainty, with a principal contribution stemming from the Monte Carlo/lattice-physics-code solutions. The Monte Carlo uncertainty analysis eigenvalue deltas consider the limiting lattice in the nominal storage rack cell geometry.

5.4.1 Analytical Model The nominally specified material/geometry parameters for the storage rack cell were utilized, in conjunction with the limiting vanished GNF2 lattice to form the basis for the uncertainty analyses cases.

Individual parameters were perturbed accordingly, to appropriately simulate the physical changes associated with a given manufacturing uncertainty or abnormal condition.

5-8

0000-0035-732 7-SFP, Rev 2.

Non-ProprietaryVersion Figure 5 Spent Fuel Storage Rack, Eccentric Loading flV11_4'Ai ig;';ý 5;'_XFE1'e II]

f I210C' 1 AIA FFA 10,W110,541ELleIII I "o.XXE LrOELI -W,I I XIL ýA A III

_*Q Q, I X[dol 150NO IR* 1-01 L_4ýA51No W,AlII III11261 "ýC x Lm Al 11*15119 1Arn.-

Eqc'Ldý1 I111 110, I F ArjAK,_'1A 110, 11 I11141ý11*1 TA AXE"All _-All I I I,,.1 L'I *I,-- 'A 54ý"0' 111 16

ýý7 A ýZ!

IW,x x w re,rA w.le,.r 1,X 1.

1 III_'.]*'.]ww ACLZL.ýw III 1111,011011,01WW'AIC911 'LIVIII 119FA 5w 50" w W, I I U'4156' 5wXN wVIJ 5-9

0000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion 5.4.2 Results 5.4.2.1 Fuel enrichment The uncertainty in the fuel enrichment analysis simulates the storage of the most reactive lattice, including flow channels, with nominal specifications of the associated storage rack material/geometry parameters, ((

)). This increase in enrichment results in an incremental reactivity uncertainty of

+0.0014 Ak.

5.4.2.2 Stainless Steel Liner Thickness The uncertainty in the stainless steel box thickness analysis simulates the storage of the most reactive lattice, including flow channels, with nominal specifications of the associated storage rack material/geometry parameters, with a 10% (0.0075") decrease associated with a nominally specified box thickness of 0.075 inches. This stainless steel thickness reduction resulted in essentially no statistically meaningful change in the system reactivity and is consistent with the results for GE14 fuel.

5.4.2.3 Lattice Spacing The uncertainty in the lattice spacing analysis simulates the storage of the most reactive lattice, including flow channels, with nominal specifications of the associated storage rack material/geometry parameters, with a 1 mm (0.1 cm) decrease in the nominally specified storage rack cell dimension of 15.9512 cm.

Decreasing the cell pitch causes an associated increase of +0.0082 Ak in the modeled rack reactivity.

The associated decrease in the storage rack cell pitch results in a strong positive eigenvalue contribution, and a 1 mm change was chosen to be consistent with the GE14 analysis performed for the racks.

5.5 Boraflex Degradation AEA Technology performed a detailed Boraflex degradation analysis for the PBAPS racks in July of 2000 (Reference 10) which was an update to an earlier analysis performed in 1996 (Reference 17). The Reference 10 analysis assumed a generic 7x7 BWR spent fuel lattice and evaluated the reactivity effects of both uniform B-10 density reduction as well as the reactivity effects of gaps in the Boraflex panels of various sizes. Based on this work, a reactivity bias penalty of +0.0131 Ak was recommended by Exelon for an assumed 10% B-10 uniform atom density reduction in all panels and 10 cm randomly distributed axial gaps in all panels. For this updated analysis, Exelon assessed a reactivity bias penalty of +0.0258 Ak for an assumed 35% B-10 uniform Boraflex panel-thinning penalty (Reference 18).

5-10

0000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion In order to confirm the applicability of the +0.0131 Ak and the +0.0258 Ak reactivity penalty to the 10xl0 GNF2 lattices used in this report, the reactivity effects of 10% to 35% uniform B-10 atom density reduction in all Boraflex panels was evaluated and compared to the published results in Table 1 of Reference 10 and Table 4 of Reference 17. Figure 6 show a comparison of these results.

Figure 6 NETco. Reactivity Comparison and Boraflex Degradation Penalty 2.50 --------.-- AEA7x7)1996 -

-*---AEA(7x7)2000 I o 2.00 ........ -- - GNF2 (10xlO ) .... ..------...---

• i I I I

'U Co III C.S

• 1.00...

(D 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0

% B-1 0 Reduction 5-11

0000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion As can be seen from the results on Figure 6, the reactivity increase (given in %Ak/k) for all three lattices shows fairly good agreement considering the differences in the initial 235U enrichments, Gadolinium loadings, number/placement of Gadolinium bearing rods, lattice geometries (i.e., 7x7 versus IWxlO),

computer codes (MONK versus MCNP) and nuclear cross-sections used to perform the analyses.

Based on this comparison, it will be assumed that the reactivity increase from randomly distributed 10 cm axial gaps in all Boraflex panels for the IOx 10 GNF2 lattice would be similar to that of the 7x7 lattice analyzed by AEA in Reference 10 (+0.0059 Ak). Therefore, the +0.0258 Ak reactivity penalty for a 10%

B-10 density reduction and 10 cm randomly distributed gaps in all panels provides a conservative representation of the Boraflex degradation and will be applied directly to the results for GNF2 (as a bias) with no further qualification. This is summarized in Table 5.

5-12

0000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion Table 5 PBAPS Criticality Safety Analyses Summary Nominal eigenvalue plus biases Reference rack eigenvalue' 0.9 018

• • • • Monte Carlo calculational bias (see Sec. 2.3) 0.0 020

• Non-channel loading bias (See Sec. 5.3.2.5) (-0.0C)30)

Boraflex degradation bias 0.02 581 k(nom+bias) 0.%9296 Uncertainties Lattice Item Aku Reactivity Report Section MCNPO1A relative error(95% confidence) Aku 1 3.4 MCNPO IA/ENDF-V (95/95 uncertainty)

• • • MCNPOIAiTGBLA06A cold k-. --- 2.3/4.3 Eccentric loading (diagonal off-set) AkU3 0.0012 5.3.2.4 Fuel enrichment tolerance Aku4 (( 5.4.2.1 Lattice Spacing (minimum) AkU5 0.0082 5.4.2.3

• SS box thickness tolerance Aku 6 (-0.0001) 5.4.2.2

• • Cold water (20'C -W40 C) Aku 8 0.0017 5.3.2.1 Total --- 0.0094 ---

Maximum reactivity (kmax) 0.9390

[ k(nom+bias) + uncertainties I

+NOTE: This value differs slightly from the result in Table 4 since the reference case simulation was run using a larger number of neutron histories to reduce the statistical uncertainty.

• NOTE: Negative effects are not included in the roll-up or bias and are indicated parenthetically above.

• • NOTE: All calculations were performed assuming 20'C pool water except for the Aku8 case.

• • • NOTE: Although there is not a k-uncertainty (ku) term associated with this bias, the 0.0034 is included in the statistical roll-up specified above.

• • • • NOTE: Treated as a bias. Not included in bias or statistical roll-up 5-13

0000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion

6.

SUMMARY

The Westinghouse Boraflex spent fuel storage racks at the PBAPS were analyzed with a bounding GNF2 10xl0 lattice fuel design with a peak, cold, in-core k-. of 1.318 at the 70 percent void history design basis. The analyses were performed with the MCNPO1A Monte Carlo neutron transport program. All storage rack conditions were analyzed assuming the GNF2 fuel lattice geometry with uniform enrichment distributions, explicit burnable absorbers and fission products for the spent fuel racks. The spent fuel storage racks were analyzed with a limiting lattice of 4.90 wt% 235U with 2 wt% gadolinium concentrations in 13 rods on the current GNF2 bundle design, corresponding to a MCNPO1A in-core k-of 1.321 anda spent fuel MNCPO1A in-rack kma. of 0.901. All analyses resulted in a storage rack kmax <

0.95 for normal and abnormal configurations, including all biases and uncertainties. Note that an in-core peak eigenvalue (k-) of 1.318 is a production lattice physics (TGBLA06A) computed value. Evaluated in conjunction with the 4.90 wt% 235U lattice average enrichment limit and the GNF2 Design Basis descriptions, this result establishes the GNF2 spent fuel storage rack design limit for the PBAPS Boraflex spent fuel storage racks.

Additional data showed a decreasing trend in peak in-core and in-rack neutron multiplication values with increasing gadolinium pellet concentrations in zero and 70 percent void history cases. Previously simulated in-core eigenvalues produced a maximum in-core (k-) of 1.359 with uniform fuel lattice enrichments of 4.90 wt % 23'U and 1.00 wt% gadolinium concentrations in 13 specified rods. Further lattice simulations resulted in a maximum in-core eigenvalue (k-) of 1.345 and 1.318 at zero and 70 percent voids respectively with uniform fuel lattice enrichments of 4.90 wt% 235U and 2.00 wt%

gadolinium concentrations in 13 specified rods.

Moreover, in-rack criticality calculations were also effected with the increase in gadolinium concentration of the new GNF2 fuel bundle specifications. Previous MCNP simulations showed a peak in-rack eigenvalue (kin-rack ) of 0.919 with uniform enrichments of 4.90 wt% 23 5 U and 1.00 wt%

gadolinium concentrations. Increasing neutron absorber concentration, MCNP in-rack calculations estimated a peak in-rack eigenvalue (kin-rack) of 0.916 and 0.901 at zero and 70 percent void histories respectively. Lattice specifications remained consistent with simulated uniform fuel enrichments of 4.90 wt% 235 U and 2.00 wt% gadolinium concentrations in 13 rods. A higher peak in-rack eigenvalue (kin-rack) is expected for the zero percent void history as a result of greater in-core moderation that depletes the gadolinium at a greater rate relative to the 70% in-channel void condition.

A review of the peak cold in-core reactivity of GNF bundles previously used over the life of Peach Bottom Units 2 and 3 indicates that this GNF2 lattice provides a bounding rack reactivity. The peak cold in-core reactivity for previous lattice designs is contained the Table 6. All previous lattice designs meet the 1.318 in-core limit.

6-1

0000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion Table 6 Peak Cold Uncontrolled Lattice Reactivity Plant: Peach Bottom 2 Lattice Lattice In-core Type Number K-infinity 8x8 694 9x9 1904

)) 1OxlO 7257 Plant: Peach Bottom 3 8x8 1180 LL 9x9 2030

] OxlO 5055 6-2

0000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion

7. REFERENCES

1. Westinghouse Electric Company, "Design Report of High Density Spent Fuel Storage Racks for Philadelphia Electric Company Peach Bottom Atomic Power Station Units 2 & 3," WNEP-8542, June 18, 1985.

2. J.F.Briesmeister, "MCNP - A General Monte Carlo N-Particle Transport Code, Version 4A,"

LA-12625-M Manual, Los Alamos National Laboratory, (1993).

3. USNRC Regulatory Guide 1.13, "Spent Fuel Storage Facility Design Basis," Rev. 2, March 2007.

4. ANSI/ANS-57.2-1983, "Design Requirements for Light Water Reactor Spent Fuel Storage Facilities at Nuclear Power Plants," October 1983.

5. Reference deleted.

6. International Criticality Safety Benchmark Evaluation Project (ICSBEP), LEU-COMP-THERM-00 1 - 063, Benchmark experiments, September 2003.

7. Reference deleted.

8. Reference deleted.

9. General Design Criterion 61, "Fuel Storage and Handling Criteria for Nuclear Power Plants," of Appendix A, "General Design Criteria for Nuclear Power Plants," to 10 CFR 50, "Licensing of Production and Utilization Facilities."

10. AEA Technology Report, "Criticality Assessment of the Peach Bottom Spent Fuel Ponds with Degraded Boraflex Panels," AEAT/R/NS/0084, Rev. 1, July 2000.

11. Reference deleted.

12. TGBLA Version 06 Letter from S.A. Richards (NRC) to G.A. Watford (GE), "Amendment 26 to GE Licensing Topical Report NEDE-2401 1-P-A, GESTAR HI Implementing Improved GE Steady-State Methods," (TAC NO. MA648 1), November 10, 1999.

13. R.D. Carter, G.R. Kiel and K.R. Ridgway, Criticality Handbook Volume I, Atlantic Richfield Hanford Co. Report, ARH-600, 1968.

14. Reference deleted.

15. Reference deleted.

16. Reference deleted.

17. AEA Technology Report, "An Assessment of the Possible Effects of Boraflex Degradation on K-Effective for the Peach Bottom Storage Pools," AEAT-079 1, October 1996.

18. NETco. Reactivity comparison and Boraflex degradation penalty values. Net-246-02.

7-t

O000-0035-7327-SFP, Rev 2.

Non-ProprietaryVersion

19. NEDC-32868P, "GEl4 Compliance with Amendment 22 of NEDE-2401 I-P-A (GESTAR If),"

Rev. 2, September 2007.

20. NEDE-30130-P-A, "Steady-State Nuclear Methods," April 1985.

21. NEDE-2401 I-P-A-16, "General Electric Standard Application for Reactor Fuel (GESTAR I),"

October 2007.

7-2

0000-0035-7327-SFP, Rev. 2 Non-ProprietaryVersion Appendix A INFINITE LATTICE BENCHMARKS The following case figures and descriptions were among those used in evaluating the PBAPS Boraflex spent fuel storage facilities with GNF2 fuel. The referenced lattices were burned with TGBLA06A based on the standard hot uncontrolled depletions, with appropriate in-core cold, xenon-free restart cases performed to identify the cold, exposure-dependent peak eigenvalues. The cold, exposure-dependent eigenvalues were found to be at their peak when the zero void history depletions were used.

Subsequently, the burned lattices were evaluated in the nominal PBAPS Boraflex spent fuel storage rack configuration at the respective peak-eigenvalue statepoints. Consistent with the methodology herein employed, the infinite lattice k- are compared with the nominal storage rack keff. The results are plotted in Figure A 1 and included in Table A 1.

Revaluated lattices executed the cold, exposure dependent peak eigenvalues at 70% void histories. This would ensure proper representations of core and fuel like behavior.

Lattice Peak 543212 GNF2-(( 1] is the design basis GNF2 lattice design (T6) to satisfy the spent fuel storage rack eigenvalue criteria, in conjunction with the design parameters and calculational assumptions applied. The accompanying graph (Figure A 1) demonstrates the constant slope relationship between in-core and in-rack eigenvalues within a product line/lattice design family. The slight offset between the plotted vanished and dominant reactivity lines indicates a difference in rack efficiency ratios between these lattice designs, within the same product line.

A-1

0000-0035-7327-SFP, Rev. 2 Non-ProprietaryVersion Figure A 1. PBAPS Boraflex Rack In-Core vs. In-Rack Eigenvalues PBAPS Boraflex Rack In-Core vs. In-Rack Eigenvalues 0.95 ,

y=0.7198x - 0.0547 R 2 =0.99571*"

y = 0,715ýx - 0.0493

,R=b.9956 I

• ,, 0.85 -- - -- - - - - -I - - - - - -I-- - - - -

0.80 Dominant (GNF2) 0.75 - ..... Vanished (GNF2)

-Linear (Vanished (GNF2))

0.70 - Linear (Dominant (GNF2))

1.14 1.19 1.24 1.29 1.34 1.39 In-Core k-inf NOTE: The Figure above is the simulated TGBLA peak, cold in-core zero percent void history eigenvalues for in-rack (k-) eigenvalues as a function of in-core criticality. Lattice specifications remained consistent with specified uniform fuel enrichments of 4.90 wt% 235U. Both bundles contained either 2.00 wt% or 1.00 wt% gadolinium concentrations in 13 rods. New dominant trendline shows a linear slope of y = 0.7198x - 0.0547 similarly to old linear trendline behavior of y = 0.7155x - 0.0493.

A-2

0000-0035-7327-SFP, Rev. 2 Non-ProprietaryVersion Table A 1 - GNF2 Results Lattice ID j TGBLA06A k-L MCNP01A kom

.4 t 4 t 4 +

1] 11 NOTE: In the table above, the TGBLA06A k-, values are the individual lattice eigenvalues computed in the in-core, reflecting boundary condition geometry. The MCNP01A knom values are the associated Monte Carlo storage rack eigenvalues at 20'C, computed with the in-rack reflecting boundary condition geometry, without biases and uncertainties included.

A-3

O000-0035-7327-SFP, Rev. 2 Non-ProprietaryVersion Appendix B IN-CORE K_ CRITERIA METHODOLOGY The basic criterion associated with the storage of irradiated (spent) and new fuel is that the effective multiplication factor of fuel stored under normal conditions will be _<0.90 for regular (low) density storage racks and _<0.95 for high-density storage racks. Abnormal conditions are limited to klff < 0.95 for both low and high density fuel storage racks. These storage rack criteria will be satisfied if the uncontrolled lattice k- calculated in the normal reactor core configuration meets certain storage rack specific k- criterion limits.

Normal and abnormal spent fuel storage rack configurations shall be evaluated with two dimensional geometry models. The analysis conditions and assumptions must in compliance with the requirements contained in the USNRC Regulatory Guide 1.13, Spent Fuel Storage Facility Design Basis, Rev. 2, (Reference 3) and ANSIIANS-57.2-1983, Design Requirements for Light Water Reactor Spent Fuel Storage Facilities at Nuclear Power Plants (Reference 4).

The in-core k- criterion method relies on a well-characterized relationship between infinite lattice k- (in-core) for a given fuel design (i.e., GNF2) and a specific spent fuel storage rack k- (in-rack) containing that fuel. The use of an infinite lattice k- criterion for demonstrating compliance to fuel storage criticality criteria has been used for all GEH-supplied spent fuel storage racks, and is currently used for re-rack designs at a number of plants where earlier criteria have become limiting for modem fuel designs.

NEDC-32868P, Rev. 2, September, 2007, "GE 14 Compliance with Amendment 22 of NEDE-2401 1-P-A (GESTAR II)" (Reference 19) presents generic information relative to the GNF2 fuel design and analyses of GE BWRs for which GNF provides fuel. The report consists of a description of the fuel licensing acceptance criteria as specified by Amendment 22 of GESTAR (General Electric Standard Application for Reactor Fuel) and the basis for generic compliance of GE fuel designs with those criteria.

Section 2.3.7, Fuel Storage, of NEDC-32868P, referencing Subsection 1.1.3.G of Amendment 22 states, "The effective multiplication factor for fuel designs stored under normal and abnormal conditions shall be shown to meet fuel storage limits by demonstrating that the peak uncontrolled lattice k-infinity calculated in a normal reactor core configuration meets the limits provided in Section 3 (of GESTAR-H) for GE-designed racks." This approved methodology has been applied to storage racks of similar design to those GE-designed racks.

The analysis performed to calculate the lattice k1 to confirm compliance with the above criterion uses the lattice physics portion of the methods described in Subsection 2.1.1 of "Steady-State Nuclear Methods,"

NEDE-30130-P-A (Reference 20). These NRC-approved lattice physics models are encoded into the TGBLA Engineering Computer Program (ECP). One of the outputs of TGBLA solution is the lattice k-of a specific nuclear design for a given set of input state parameters (void fraction, control state, fuel temperature). A description of the requirements and the analytical process to calculate the fuel storage reactivity requirements is contained in Section 3.5 of NEDE-2401 1-P-A-16 (Reference 21).

B-1

0000-0035-7327-SFP, Rev. 2 Non-ProprietaryVersion Compliance of GNF2 fuel with specified 1L limits will be confirmed for each GNF2 lattice as part of the bundle design process. Documentation that this has been met will be contained in the fuel design information report which defines the maximum lattice kL for each bundle nuclear design.

PROCESS FOR VALIDATING SPECIFICGNF2 ASSEMBLY DESIGNS ARE ACCEPTABLE FOR STORAGE IN THE PBAPS BORAFLEX SPENT FUEL STORAGE RACKS

1. Identify the unique lattices in each assembly design.

2. Determine the maximum cold 68°F uncontrolled k-. results from standard TGBLA06 lattice physics calculations for each unique lattice in the assembly.

3. Ensure that the k- values obtained from Step 2 for each lattice are less than the kL limit of 1.318.

Note: kL, limit of 1.318 is based on TGBLA generated GNF2 lattices with 4.90 wt% 235U enrichments and 2 wt% gadolinium concentrations.

Note that lattice physics codes other than TGBLA06 may be used to perform the evaluation, subject to a determination of an appropriate k- limit having been determined for the particular lattice physics code to be used.

B-2

O000-0035-7327-SFP, Rev. 2 Non-ProprietaryVersion Appendix C CRITICALITY ANALYSIS FOR FUEL INSPECTION The criticality analyses that support the fuel inspection and reconstitution activities are based on calculations using Monte Carlo transport theory methodology.

The results of these analyses are as follows:

((

1]

The intact fuel rod analysis was based on 5% enriched uranium. A triangular fuel rod array with optimum fuel rod pitch was conservatively assumed. The water density was varied to determine the optimum moderator conditions for each fuel rod diameter. No credit was taken for natural uranium end-segments, gadolinia burnable poisons, fuel rod depletion, nor any fission product inventory.

The maximum pellet free mass calculation was also based on 5% 235U. An optimum spherical heterogeneous mass of pellets and moderator is assumed. Similar to the fuel rod calculations, no credit was taken for natural uranium end-segments, gadolinia burnable poisons, fuel rod depletion, nor any fission product inventory.

The GNFA fuel examination procedure conservatively limits the number of sound rods outside of a bundle loaded in the fuel preparation machine, at any time, to 31. In addition, these procedures limit the number of failed rods outside the bundle to three. In this way, it is assured that the maximum amount of material in an uncontrolled geometry is always maintained subcritical.

C-t