ML20207G860

to Criticality Safety Analyses for Vogtle Electric Generating Plant Spent Fuel Storage Racks

ML20207G860
Person / Time
Site:	Vogtle
Issue date:	08/12/1988
From:	GEORGIA POWER CO.
To:
Shared Package
ML20207G842	List: ML20207G836 ML20207G860
References
NUDOCS 8808240258
Download: ML20207G860 (34)

Text

_ _ _

Criticality Safety Analyses For Vogtle Electric Generating Plant Spent Fuel Storage Racks 1

l l

l l

l l

l Rev. 1 8/12/88 8808240258 880842 PDR ADOCK 05000425 A PDC i

. \

TABLE OF CONTENTS Pace 4.0 CRITICALITY SAFETY ANALYSES........................... 4-1 4.1 DESIGN BASES..................................... 4-1 4.2

SUMMARY

OF CRITICALITY ANALYSES.................. 4-2 4.2.1 Normal Operating Conditions............ 4-2 4.2.2 Abnormal and Accident Conditions....... 4-5 4.2.3 New Fuel Storage in Dry Condition...... 4-6 4.3 REFERENCE I'UEL STORAGE CELL . . . . . . . . . . . . . . . . . . . . . . 4 -6 4.3.1 Reference Fuel Assembly................ 4-6 4.3.2 Fuel Storage Cells..................... 4-7 4.4 ANALYTICAL !ETHODOLOGY 4-10 4.5 CRITICALITY ANALYSES AND TOLERANCE 4-14 VARIATICHS........................................

4.5.1 Nominal Design Case.................... 4-14 4.5.2 Uncertainties Due to Manufacturin 4 Tolerances.......................g .......

4.5.2.1 Boron Loading Variation................ 4-14 4.5.2.2 Storage Cell Lattice Pitch Variation... 4-15 4.5.2.3 Boraflex Width Tolerance Variation..... 4-16 4.5.2.4 Boraflex Integrity..................... 4-16 4.5.2.5 Stainless Steel Thickness Tolerances... 4-17 4.5.2.6 Fuel Enrichment and Density Variation.. 4-17

. 4.5.2.7 Eccentric Positioning of Fuel Assembly 4-17 in Storage Rack.........................

4.5.3 Reactivity Effects of Boraflex Axial 4-18 Length 4.5.4 Reactivity Effects Due to Retainer 4-18 1 Spring 4.6 ABNORMAL AND ACCIDENT CONDITIONS 4-19 4.6.1 Temperature and Water Density Effects.. 4-19 4.6.2 Dropped Fuel Assembly.................. 4-20 4.6.3 Abnormal Location of a Fuel Assembly... 4-20 4.6.4 Seismic Event........................... 4-20 4.7 FUEL STORAGE UNDER DRY CONDITIONS................ 4-21 REFERENCES 4-22 APPENDIX A - Benchmark Calculations Rev. 1 8/12/88

LIST OF TABLES No.

Pace 4.1

SUMMARY

OF CRITICALITY SAFETY ANALYSES 4-4 4.2 REACTIVITY EFFECTS OF ABNORMAL AND ACCIDENT 4-5 CONDITIONS 4.3 FUEL ASSEMBLY DESIGN SPECIFICATIONS 4-9 4.4 EFFECT OF TEMPERATURE AND-VOID ON CALCULATED 4-19 REACTIVITY OF STORAGE RACK (Standard and OFA Fuel). I LIST OF FIGURES EQ. Pace 4.1 VOGTLE FUEL STORAGE CELL 4-8 l1 Rev. 1 '8/12/88

l l

- 4.0- CRITICALITY SAFETY AllALYSES 4 .' 1 DESIGli BASES The high density spent fuel storage racks for the

-Vogtle Flectric Generating Plant are designed to assure-that the neutron - multiplication factor (keff) is equal to or less than-i 0.95 with the racks fully loaded with fuel of the highest anticipated reactivity, and flooded with unborated - water at - a 3 temperature corresponding to the highest reactivity. The maximum '

calculated reactivity includes a calculational bias, a margin for  ;

uncertainty in reactivity calculations and in mechanical i

tolerances. The uncertainties in reactivity calculations and mechanical tolerances are statistically combined such that the~ l true keff will be equal to or less than 0.95 with a 95%

probability at a 95% confidence level.

l l

Applicable codes, standards, and regulations, or pertinent sections thereof, include the following:

O General Design Criterion 62, Prevention of Criticality in Fuel Storage and Handling.

O US!!RC Standard Review Plan, 11UREG-0800, Section 9.1.1, l New Fuel Storage, and Section 9.1.2, Spent Fuel Storage.

O US11RC letter of April 14, 1978, to all Power Reactor Licensees -

OT Position for Review and Acceptance of Spent Fuel Storage and Handling Applications, including modification letter dated January 18, 1979.

O US11RC Regulatory Guide 1.13, Spent Fuel Storage Facility Design Basis, Rev. 2 (proposed), December, 1981.

O U S 11 R C Regulatory Guide 3.41, Validation of Calculational Methods for 11uclear Criticality Safety (and related A11SI 1116.9-1975).

O A11SI/ Alls-57.2-1983, Design Requirements for Light Water-Reactor Spent Fuel Storage Facilities at 11uclear Power Plants.

4-1

4 0 .

O ANSI N210-1976, Design Objectives for Light Water Reactor Spent ~ Fuel. Storage-Facilities at Nuclear Power-Plants.

O AUS-8.17-1984, Criticality Safety Criteria for the Handling, Storage and Transportation of- LWR Fuel Outside Reactors.

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

0 ~ Moderator is pure, unborated water at a temperature within the design basis range corresponding to the highest reactivity.

O Lattice of storage racks is assumed infinite in all directions, i.e., no credit is taken for axial or radial neutron leakage (except in the assessment of certain abnormal / accident conditions).

O Neutron absorption in minor structural members is neglected, i.e., spacer grids are replaced by water.

~

O The fuel of highest anticipated' reactivity is assumed, i.e., fresh un-burned fuel of 4.55% enrichment, in the Westinghouse geometries.

17x17 standard and optimized fuel assembly 1 1 i

1 4.2 SUlG!ARY OF CRITICALITY ANALYSES 4.2.1 Normal Ocerating Conditions The Vogtle Plant spent fuel storage racks have been analyzed for criticality safety using conservative estimates of  ;

the uncertainty in enrichment and in B-10 loading in the Boraflex poison material and using as-built measurements of the flux-trap 1 water gap between storage cells. Based upon Westinghouse standard fuel assemblies of 4.55% enrichment, (57.48 grams U-235 per axial centimeter), the maximum infinite multiplication factor (km) is 0.943 (95% probability at the 95% confidence level) for Rev. 1 8/12/88 4-2

the rack module with the most limiting (i.e., smallest) measured average water-gap, including all known uncertainties. Table 4.1 summarizes results of the criticality safety analyses for Rack Module A-5, which bounds all other modules.

{

A' change in manufacturing procedures enabled Rack module B-2 and subsequent modules to be manufactured with larger water-gaps and s

1 closer tolerances. 'Therefore, Rack module B-2, shown in Table 4.1, constitutes a bounding case'for all rack modules except A-1, A-2, A-4, and B-1 (which are bounded by the analysis of Rack A- s 5). The maximum infinite multiplication f actor - (km, 95%/95%,

including uncertainties)-for Rack B-2 is 0.932 with standard fuel and 0.941 with optimized fuel (both evaluated at 4.55% U-235 enrichment).

4-3 Rev. 1 8/12/88

I Table 4.1

SUMMARY

OF CRITICALITY SAFETY ANALYSES Rack A-5 Rack B-2 T.estinghouse type fuelu Standard. Standard 'QEA Reference km (AMPX-KENO) 0.9075 0.9075_ 0.9162 Calculational bias, Ak 0.0106 0.0106 0.0106 Meas. water-gap effect, Ak 0.0127 0.0016 0.0034 SS spring effect, Ak 0.0018 0.0018 0.0018 Total 0.9326 0.9215 0.9320 Uncertainties, Ak (Note 1) 10.0107 t0.0107 0.0089 Maximum km 0.9433 0.9322 0.9408 (Note 1)

Detail of Uncertainties Standard OFA i Calculational i 0.0071 i 0.0044 Bias 0.0048 i 0.0048 B-10-concentration 0.0026 i 0.0020 Boraflex thickness 1 0.0031. 0.0023 Boraflex width i 0.0030 t 0.0017 Inner box dimension 1 0.0010  ! 0.0012 Gap measurement accuracy i 0.0028 0.0034 SS thickness t 0.0002 i 0.0005 Fuel enrichment 0.0019 i 0.0020 Fuel density 0.0018 0.0024 Eccentric assembly position < 0.0001 < 0.0001 Statistical combination of 0.0107 1 0.0088 uncertainties (Square root of sum of squares) 4-4 Rev. 1 8/12/S8

l l

\

4.2.2 Abnormal and Accident Conditions  !

Although credit for the solu:1.e poison normall'y present in the spent fuel pool water is permitted under abnormal or j accident conditions,* most abnormal or accident conditions will l not result in exceeding the limiting reactivity (keff of 0.95) even in the absence of soluble poison. The effects on reactivity l of credible abnormal and accident conditions are summarized in Table 4.2 below. Of these abnormal / accident conditions, only two {

have the potential for a more than negligible positive reactivity l effect. l Table 4.2  !

l REACTIVITY EFFECTS OF ABilORMAL AllD ACCIDEllT COllDITIOllS j l

Accident / Abnormal Conditions Reactivity Effect  !

Temperature increase llegative  :

I Void (boiling) llegative Assembly dropped on top of rack 11egligible Lateral rack module movement (seismic) Positive Misplacement of a fuel assembly Positive Double contingency principle of A11SI 1116.1-1975, as specified in the April 14, 1978 IIRC letter (Section 1.2).and implied in the proposed revision (draft) to Reg. Guide 1.13 (Section 1.4, Appendix A).

Either a major seismic event or the misplacement of a fresh fuel assembly outside and adjacent to a rack module have the potential for exceeding the limiting reactivity should there be a concurrent and independent accident condition resulting in the loss of all soluble poison. Administrative procedures will assure the presence of soluble poison after an earthquake and 4-5

during fuel handling operations and preclude the' possibility of the simultaneous occurrence. of two independent accident conditions.

4.2.3 New Fuel Storace in Dry Condition New fuel is normally stored in the dry condition with a very low multiplication factor. The misplacement of a . fuel assembly during fuel handling operations has no criticality consequences under these normally dry conditions. NRC guidelines (SRP 9.1.1) require the evaluation of criticality safety under flooded conditions and for hypothetical low-density moderation.

The double contingency principle of ANSI N16.1-1975 (invoked by the NRC April 14, 1978 letter) eliminates the i.ecessity of considering independent and concurrent accident conditions. .

Therefore, the conditions specified in SRP 9.1.1 are the accident conditions evaluated.

The storage rack in the ficoded condition'is the reference design basis case previously evaluated. With Boraflex absorber between assemblies, conditions do not exist for the appearance of ,

a peak in reactivity at low moderator densities, and the fully flooded condition corresponds to the highest reactivity (optimur.

moderation). Calculations at low moderator densities (5% to 15%)

confirm the very low reactivities. At the 10% water density where low density "optimum" moderation might otherwise be y expected, the infinite multiplication factor is not more than 0.56.

4.3 REFERENCE FUEL STORAGE CELL 4.3.1 Reference Fuel Assembly Tha design basis fuel assembly, illustrated in Fig.

4.1, is a 17 x 17 array of fuel rods with 25 rods replaced by 24 control rod guide tubes and 1 instrument thimble. Table 4.3 summarizes the design specifications and the range of significant variations. Independent calculations with Westinghouse standard l1 4 -6 Rev. 1 8/12/88

i and with OFA (optimized) fuel assemblies (see Table 4 .1 ) ,- I confirmed that the OFA assembly exhibited the highest reactivity.

4.3.2 Fuel Storace Cells The nominal spent fuel storage cell used for the criticality analyses is shown in Fig. 4.1. The rack is composed of Boraflex absorber material sandwiched between an 8.75 ' inch I.D., 0.075-inch thick inner stainless steel box, and a 0.020-inch outer stainless steel coverplate. The fuel assemblies are centrally located in each storage cell on a nominal lattice spacing of 10.40 inches in one direction and 10.58 inches in the 1 other direction. Stainless steel tabs connect one storage cell box to another in~a rigid structure and define an outer water space between boxes. This outer water sp. ace constitutes a flux-trap between the two -Boraflex absorber sheets that are each essentiall" opaque (black) to thermal neutrons. The Boraflex absorber has a thickness of 0.075 1;0.007 inch and a nominal B-10 areal density of 0.0238 gram per cm2 . , Measurements of the as-built water gaps were used in the cri.ticality evaluation for greater accuracy. Thin (0.020 inch) stainless cover plate sheets are used to provide support for the Boraflex sheets. It was assumed that stainless steel spring retainer springs (elongated 1 "C"

shape) are used to backup the cover plates in all modules for calculational purposes. '

4-7 L

Rev. 1 8/12/88 1

1 I

l . .

l l

l

--!------ ,,,,,o 3. ,on,,t,x -

t 0.07510.007" THK 0.0238 810/cm2  !

I I

l

!^I $'O000000000000000

- I n .

! 00000000000000000

0.m- l l 00000000000000000

i iANGLEi 00090000000009000 i j'!;&o

2 li t

00000000000000000 : 13

' OO900SOO900900900 000000000000000006 n TuNEn diNo ! 00000000000000000l TYP. 4 SIDES  : 00400400900900900

i i( ,

j 00000000000000000 j i 00000000000000000 W ten cie 3 00000000eO0900900 ::

i.*0 N s i l s

00000000000000000 :

! OOO90000000009000!

.280 E - W I

t i 0000(

00000000000000000 i

'5 90 0 i ,

I D O O O G v. 73,0.03. oox iovvvvvvJOOO - -

l ._

' 0.075" SS BOX 10.58 N,S DIRECTION

'~

10.40 E W DIRECTION '

.-.-......................................~.;.~.....................;...... p NOT TO SCALE '

FIGURE 4.1 VOGTLE PLANT FUEL. STORAGE CELL

ip Table 4.3 FUEL ASSEMBLY DESIGN SPECIFICATIONS FUEL POD DATA STANDARD OFA Outside diameter, in. 0.374 0.360 Cladding thickness, in. 0.0225 0.0225 Cladding inside diameter, in. 0.329 0.315 Cladding material Zr-4 Zr-4 i Pellet density, % T.D. 95.0 95.0 Pellet diameter, in. 0.3225 0.3088 Enrichment, wt % U-235 4.55 1 0.05 4.55 t 0.05 Stack density, gUO2 /cc 10.30 t 0.22 10.30 0.22 i Grams U-235/ axial em. 57.48 52.70 FUEL ASSEMBLY DATA '

Fuel rod array 17x17 17x17 i Number of fuel rods 264 264 Fuel rod pitch, in. 0.496 0.496  ;

Number of control rod guide 24 24 tubes '

Guide thimbles, O.D., in. 0.484 0.474 Guide thimbles, I.D., in. 0.448 0.442 j

.. .I Number of instrument thimbles 1 1 Instrument thimble, 0.484 0.474 0.D., in.

Inatrument thimble, 0.448 0.442 I.D., in.

l l

l 4-9 Rev. 1 S/12/88

l 4.4 AllALYTICAL 14ETHODOLOGY In the fuel rack analyses, criticality analyses of the high density spent fuel storage racks were performed with the AMPX-KEllO computer package (Refs. 1 and 2), using the 27-group SCALE

• cross-section library (Ref.. 3) with the NITAWL (ref. 1) subroutine for U-238 resonance shielding effects (11ordheim integral trectment). Benchmark calculations are presented in Appendix A and indicate a bias of 0.0106 0.0048 (95%/95%). In the geometric model used in KEllO, each fuel rod and its cladding were described explicitly and reflecting boundary conditionc (zero neutron current) were used i.i the axial direction and at the centerline of the water-gap between storage cells. These boundary conditions have the effect of creating an infinite array of storage cells in all directions.

The CASMO-2E computer code (Refs. 4, 5 and 6), a two-dimensional multigroup transport theory code for fuel assemblies, has also been benchmarked (Appendix A) and was used both for verificatier. calculations and as a means of evaluating small reactivity increments associated with manufacturing 31erances.

CASMO-2E benchmarking resalted in a calculational bias of 0.0013 0.0018 (95%/95%). However, limitations in the geometry options available in CASPJ-2E required minor approximations in the geometric description (e.g. in the description of the Boraflex absorber and in the use of an average water-gap thickness) which apparently contributes to a small over-prediction in the absolute SCALE is an acronym for Standardized Computer Analysis for Licensing Evaluation, a standard cross-section set developed by ORIIL for the US!1RC.

4-10

value of the CASMO cell infinite multiplication factor. Two group diffusion theory constants were edited in the. output of CASMO-2E and used in the one dimensional SNEID* diffusion theory routine to evaluate reactivity effects of the Boraflex ~ axial length and variation in water gaps.

A third independent method of criticality analysis, utilizing diffusion / blackness theory, was also used with the OFA l1 fuel for additional confidence in results of the primary calculational model and methods, although no reliance for l1 ;

criticality safety is placed on the reactivity value from the diffusion / blackness theory technique. This technique, however, is used for auxiliary calculations of the small incremental reactivity effect of eccentric fuel positioning that would otherwise be lost in normal KENO statistical variations, or would l be inconsistent with CASMO-2E geometry limitations. Cross sections for the diffusion / blackness theory calculations were l derived from the NULIF computer code (Ref. 7), supplemented by a' blackness theory routine that effectively imposes a transport theory boundary condition at the surface of the Doraflex neutron l absorber. Shielded cross-sections were then used in the spatial l diffusion theory code, PDQ-7 (Ref. 8), in two dimensions, to l I

calculate reactivities. I Comparison of the three independent methods of analysis for the reference design nominal water gap with OFA fuel resulted in l1 the following data which confirms the validity of the analytical methodology.

SNEID is a one-dimensional diffusion theory program for the microcomputer, benchmarked against one-dimensic.nal PDQ-7 calculations.

4-11 Rev. 1 8/12/88

9 l

l l

l Maximum AILalvtical Method Bias-corrected b (95%/95%) l l

AMPX-KENO 0.9268 i 0.0065 0.933 CASMO-2E 0.9420 0.0018 0.944 Diffusion-blackness 0.939 0.939 theory (PDQ-7)

These data indicate that the geometric approximations neccssary in CASMO-2E contributed to an over-prediction in b of about 1% '

Ak.

In this comparison, the calculational uncertainties listed were derived from benchmark analyses (Appendix A) anil, for KENO, includes the additional statistical variation asrociated with Monte Carlo calculations. Manufacturing uncertainties are not included in this comparison.

In evaluating the storage modules using measured as-built water-gap spacings, diffusion theory constants edited from CASMO-2E were used in one-dimensional cylindrical calculations to determine the incremental reactivity effects of the small I

variations in water-gap spacings. These small reactivity increments (Ak) were added to the nominal b from the KENO 1 calculations plus the total uncertainty from other manufacturing tolerances to obtain the maximum possible reactivity for each rack modula mecsured. Four types of calculations were made in order to assure that the analysis accurately represented the effect of the measured water-gaps.

o Two-dimensional PDQ7 calculations to demonstrate that models with (1) the two nominal water-gaps (1.46 inch in one direction and 1.28 inch in the other direction) at.d (2) with the average water-gap (1.37 inches in both directions) gave the same b, showing that in the range of the water-gap sizes employed, averaging accurately represents the reactivity of the actual geometry. Potential effects of caannel twist and bow would be averaged and therefore would 4-12 Rev. 1 8/12/88

- - . .. . _ . . - . . - - ,-. - . ~ . - - . - -.

J

e. .

p i

have a negligible reactivity effect. These could also be considered as a local subset of the evaluation of eccentric assembly position which was found to have. a negligible .

reactivity effect.

o Evaluation of the measured water-gap sizes for each-module to determine tha individual cell and the 9-cell array (3 x 3 configuration) with the minimum water-gaps. A 9-cell array represents a row of cells surrounding a central cell and is a reasonable size unit for averaging reactivity effects.

o Axial calculations with top and bottom measured gap sizes to confirm the validity of averaging the axial dimensions.

o one dimensional calculations in equivalent cylindrical geometry, with (1) the most reactive individual cell (minimum water-gap) in the center surrounded by the remaining, cells in the module at their average water-gap, (2) the mcst reactive 9-cell array (with minimum average water-gap) in the center surrounded by the I remaining cells at their average water-gap, and (3) the most reactive individual cell, surrounded . by the next row of cells (eight cells with their averaged water-gap) and a third region with the average uater gap of all remaining cells in the module.

Because of the small differences in water-gap sizes a.!d the smearing effect of neutron migration, all three of the one l dimensional calculations gave very nearly the same result. The l calculations with the 9-cell array, however, gave slightly higher and therefore more conservative km values. -Reported maximum km values represent these more conservative calculations of the  !

small reactivity increment between the nominal design water-gap 3 and the actual as-built values.

4-13 4

Rev. 1 8/12/38

7 l 1

i 4.5 CRITICALITY AllALYSES AtID TOLERANCE VARIATIO11S 4.5.1 Nominal Desian Case For - the nominal- storage cell design, the AMPX-KENO calculations resulted in a km of 0.9075 1 0.0071 and 0.9162 t 0.0044 for standard and OFA fuel respectively, including a one-sided tolerance factor for-95% probability at a 95% confidence level. Correcting for 0.0106 0.0048 Ak bias and for the measured water gaps, and combining all known manufacturing uncertainty factors, the maximum km values with standard fuel are 0.943 in Rack module A-5 and 0.932 in Rack module B-2 '(Table 1 4.1). The result for module B-2 with OFA fuel is 0.941.

4.5.2 Uncertainties Due to Manufacturina Tolerances 4.5.2.1 Boron Loadina Variation The Boraflex absorber sheets used in the storage cells are nominally 0.075-inch thick, 7.75-inches wide!and 139-inches- 'l long, with a B-10 areal density of 0.0238 g/cm2. Independent manufacturing tolerance limits are i 0.007 inch in thickness and t0.0089 g/cm3 in B-10 content. This assures that at any point where the minimum boron concentration (0!1160 gram B-10/cm3 ) and t

minimum Boraflex thickness (0.068 inch) may coincide, the boron-10 areal density will not be than 0.020 less gram /cm2 ,

Differential CASMO-2E calculations indicate that these tolerance limits result in incremental reactivity uncertainties with standard and OFA fuel, respectively, are 10.0026 and 0.0020 Ak I for boron concentration, and i 0.0031 and 10'.0023 Ak for Boraflex thickness variations.

l 1

4-14 I

Rev. 1 8/12/88

i Storace Cell Lattice Pitch variation 4.5.2.2 The design storage cell lattice . spacing between fuc1 l assemblies is 10.40 in one direction and 10.58 inches in the 1 I other direction. A decrease in storage cell lattice spacing may or may not increase reactivity depending upon other dimensional changes that may be associated with the decrease in lattice spacing. Increasing the water thickness between the fuel and the inner stainless steel box results in a small increase in reactivity. The reactivity effect of the flux-trap water thickness, however, is more significant, and decreasing the flux-trap water thickness increases reactivity. 1 The inner stainless steel box dimension, 8.750 0.03 inches, defines the inner water thicknesc between the fuel and the inside box. For the tolerance limit, the uncertainty in reactivity is ! 0.0010 and 10.0012 Ak (standard and OFA fuel) as {1 determined by diffarential CASMO-2E calculations, with km increasing as the inner stainless steel box dimension (and derivative lattice spacing) increases.

The nominal design flux-trap water thickness is 1.28 inches in one direction and 1.46 inches in the'other direction.

Actual as-built water gaps were measured and used to evaluate the reactivities of the various rack modules. The measurement accuracy was determined to be 0.04 inches, which results in an ,

uncertainty of t0.0028 for standard and of 10.0034 Ak for OFA  ;

fuels if thickness is simultaneously reduced on all four sides. 1 ,

Two independent two-dimensional calculations -

one with the nominal gap sizes (1.46 and 1.28 inches) and one with the gap l sizes averaged (1.37 inches) -

resulted in the same km values,  !

demonstrating that, in the range of the small water gaps,  !

, averaging may confidently be used.

1 l

1 5

4.5.2.3 Boraflex Width Tolerance variation The reference storage cell design uses a Boraflet sheet width of 7.75 0.063 inches. A positive increment in reactivity.

occurs for a decrease in Boraflex absorber width. For a reduction in width of the maximum tolerance, 0.063 inch, the calculated positive reactivity increments are small. However, to l

allow for radiation-induced shrinkage in width of the Boraflex and for possible small edge affects, the width tolerance was j increased from 0.063 inch to 0.25 inch corresponding to an l 1

uncertainty of i 0.0030 and i0.0017Ak for standard and OFA fuel. i 4.5.2.4 Boraflex Intecrity j The stability and integrity of the Boraflex absorber ,

material under irradiation has recently been investigated (11) j and further irradiation testing is currently underway. Available l information confirms there is no loss of boron during irradiation l t

although there is some radiation induced shrinkage. Under irradiation, Boraflex becomes a hard ceramic-like material and apparently shrinks 2 to 2-1/2 percent. At a very high radiation dose, there is evidence of a small edge deterioration. In the Vogtle racks, the Boraflex sheets are installed in a gap of sufficient size to allow unimpeded shrinkage and thereby preclude any mechanism that might cause gaps to develop.

To allow for shrinkage, the Boraflex sheets are initially 3 inches longer (approximately 2%) than would otherwise l1 be necessary. Width shrinkage is accommodated by increasing the tolerance to 10.25 inches from the nominal 0.063 inch tolerance. l1 In both cases, shrinkage would increase the boron concentration in the Boraflex although no credit is taken for this increased loading. Shrinkage in thickness would not change the B-10 areal density.

1 4-16 Rev. 1 8/12/88

-- - . - . . . . . . . . -- . . ~ . . - -

.F 4.5.2.5 Stainless Steel Thickness Tolerances The nominal stainless steel thickness is 0.075' ' ' 0.005

~

inch for the inner stainless steel box and 0.020 1 0.003 inch for the Boraflex coverplate. The-maximum positive reactivity effect of the expected stainless steel thickness tolerance variations, statistically comoined, was calculated (CASMO-2E) to be f 0.0002

/ 10.0005 Ak (standard / OFA). 1 4.5.2.6 Fuel Enrichment and Dgyusity Variation i The design maxirc'im enrichment is 4.55 t 0.05 wt%

U-235. Calculations of the sensitivity to small enrichment variations by CASMO-2E yielded a coefficient'of 0.0040 ak per.0.1 wt% U-235 as the design enrichment. For a tolerance on U-235 enrichment of i 0.05 in wt%, the uncertainty en kco is 0.0019 /

i 0.0020 Ak (standard /OFA). 1 Calculations were also made with the UO2 fuel density . 1 increased to the maximum expected value of 10.5 g/cm3 (smeared density). l'or the reference design calculations, the uncertainty in reactivity is i 0.0018 / i 0.0024 Ak (standard /OFA) over the 1 maximum expected range of UO2 densities.

i 4.5.2.7 Eccentric Positioninc of Fuel Assembly in Storace Rack The fuel assembly is assumed to be normally located in  !

the center of the storage rack cell. Calculations with the fuel assemblies assumed to be eccentrically located in the corner of the storage rack cell (four-assembly cluster at closest l approach), indicated a negligible change in reactivity as determined by differential PDQ-7 calculatiens.

Rev. 1 8/12/88 AoD -

i 4.5.3' Reactivity Effects of Boraflex Axial Lenath Based upon diffusion theory constants edited in the CASMO-2E output (reference design and.a~special case with. water replacing the Boraflex), one-dimensional axial calculations were made to evaluate the reactivity effect of reduced Botaflex axial lengths. Reduced length of the Bor.2 51ex leaves small regions of active fuel without poison at each end of the fuel assembly. The unpoisoned region at each end is referred to as "cutback".

The axial calculations used a thick (30 cm.) water '

reflector, neglecting the higher absorption of the stainless-steel structural material at the ends of the fuel assembly.

Results of the calculations showed that the keff remains less than the reference km of the storage cells until the axial reduction in Boraflex length (cutback) exceeds four inches top and bottom corresponding to a required overall Boraflex length of 136 inches. Thus, the axial neutron leakage more than compensates for the 4-inch design cutback and the reference km remains a conservative over-estimate of the true keff. In manufacturing the racks, a 4-inch cutback is used at the bottom of the rack.

However, an initial Boraflex length of 139 inches is used which 1; provides an allowance of 3 inches (approximately 2%) at the top of the racks to accommodate radiation-induced shrinkage of the Boraflex wit acut exceeding the allowable cutback.

4.5.4 Reactivity Effects due to Retainer Sorina A stainless steel spring retainer in the water gaps between cells is included in the calculation. This retainer is used to provide additional means to keep the Boraflex cover plate i sheet in place. The detailed description of this retainer is shown in Figure 4.1.

The inclusion of this stainless steel spring in the water gap increased the calculated km by 0.0018 Ak.

4-18 Rev. I 8/12/88

4.6 AB110RMAL A11D ACCIDEllT cot 1DITIO!1S 4.6.1 Temoerature and Water Density Effects The moderator temperature coefficient'of. reactivity is negative and a conservative moderator temperature of 20 0C was assumed for the reference design which assures that the true reactivity will always be lower.

Temperature effects on reactivity have been calculated and the results are shown in Table 4.4 for both standard and OFA fuel. Introducing voids in the water internal to the storage '

cell (to simulate boiling) decreased reactivity, as shown in the. 1 table. It has been shown that voids due to boiling will not occur in the outer (flux-trap) water region for the maximum bulk temperature of 170 07, i

Table 4.4 l

EFFECT OF TE!!PERATURE A11D VOID 011 CALCULATED REACTIVITY OF STORAGE RACK (Standard and OFA Fuel) l1 Case 1

Incremental Reactivity Change, Ak 20*C Reference 50*C -0.005 80'C -0.012 120*C -0.024 120'C + 20% void -0.085

~

With soluble poison present, the temperature coefficients of reactivity would differ from those inferred from the data in Table 4.4 However, the reactivities would also be substantially lower at all temperatures with soluble boron present, and the data in Table 4.4 is pertinent to the higher reactivity unborated case.

4-19 Rev. 1 8/12/88

4.6.2 Droceed Fuel Assembly Accident For a drop on top of the rack, the fuel assembly will come to rest horizontally on top of the rack with a minimum separation distance from the fuel of more than the 12 inches I sufficient to preclude neutron coupling (i.e'., an effectively infinite seperation). . Maximum expected deformation under seismic or accident conditions will not reduce the minimum - spacing between fuel assemblies to less than 12 inches. Consequently, a fuel assembly drop accident will result in only an insignificant increase in reactivity (<0.0001) due to the separation distance.

Furthermore, soluble boron in the pool water would substantially reduce the reactivity and assure that the true reactivity is always less than the limiting value for any conceivable dropped fuel accident.

4.6.3 Abnormal Location of a Fuel Assembly The abnormal location of an unirradiated fuel.assembl', 1 of 4.55% enrichment could, in the absence of soluble poison, result in exceeding the design reactivity limitation (km of 0.95). This could occur if the assembly were to be positioned outside and adjacent to a storage rack module. Soluble poison, however, is normally present in the spent fuel pool water (for which credit is permitted under these conditions) and would maintain the reactivity substantially less than the design  !

limitation.

l i

4.6.4 Seismic Event l During and after the design basis earthquake it is possible for rack modules to move so as to reduce the spacing 1 between modules to less than the design water-gap. However, the seismic analysis indicates that racks will not impact each other i or with the pool wall during an earthquake. For this condition, credit for soluble poison is permitted and will assure a keff .

less than the design limit of 0.95.

)

Rev. 1 8/12/88 1

-. _ .y - _ _

4.7 FUEL STORAGE UNDER DRY CONDITIONS For storage of new fuel in the d:ry co"dition, SRP 9.1.1 requires that the maximum reactivity shall not exceed a keff of 0.98 with fuel of the highest anticipated reactivity in place and assuming the optimum hypothetical low density moderation (i.e.,

fog or foam). With even relatively weak absorbers between fuel assemblies, however, conditions do not exist (reference 10) for a significant maxima in reactivity at low moderator densities.

AMPX-KENO calculations for an infinite array (no leakage) gave the following values:

5% water density, kw = 0.566 0.004 10% water density, km = 0.564 0.004 15% water density, km = 0.582 1 0.005 These data confirm the absence of a maxima in reactivity at low-moderator densities where the "optimum" moderation peaks would 1 otherwise be expected. L Under the accident condition of flooding, the k-infinite '

would be less than 0.943 (maximum) as previously defined (Table 1 4.1). Reducing water density reduces reactivity monotonically.

At 95% water density, k-infinity is reduced by 4.002 Ak and reduced by 0.080 Ak at 75% water density. Leakage, if included, would substantially further reduce these values.

These very low values confirm the results of Cano, et al.,

(reference 10) and demonstrate the criticality safety of the vogtle fuel storage racks under hypothetical low density moderation in conformance with SRP 9.1.1. Therefore, fresh fuel assemblies of 4.55% enrichment may be safely stcred without any criticality restrictions other than the limitation on enrichment.

4-21 Rev. 1 8/12/88

REFERENCES  ;

1. Green, Lucious, Petrie, Ford, White, Wright, "PSR-63/AMPX-1 (code package), AMPX _ Modular Code System for Generating-Coupled Multigroup Neutron-Gamma Li'.raries from ENDF/B,"

ORML-TM-3706, Oak Ridge National Laboratory, March 1976..

2. L.h. Petrie and N.F. Cross, "KENO-IV, An Improved Monte Carlo Criticality Program," ORNL-4938, Oak Ridge National Laboratory, November 1975.

3. R.M. Westfall et al., "SCALE: A Modular Code System for Performing Standaroized Computer. Analysen, for Licensing Evaluation," NUREG/CR-0200,.1979.

4. A. Ahlin, M. Edenius, H. Haggblom, "CASMO - A Fuel Assembly Bur-up Program," AE-RF-76-4158, Studsvik report (proprietary).

5. A. Ahlin and M. Edenius, "CASMO - A Fast Trar. sport Theory Depletion Code for LWR Analysis," ANS Transactions, Vol. 26,

p. 604, 1977.

3-

i. M. Edenius et al., "CASMO Benchmark Report," Studsvik/Ri' 6293, Aktiebolaget Atomenergi, 'tarch 1973. i r

7 W.A. Wittkopf, "NULIF -

Neutron Spectrum Generator, Few-Group Constant Generator and Fuel Depiction Code," BAW-426,

, The Babcock and Wi.cox Company, August 1976.

8. W.R. Cadwell, PDQ107 Reference Manual, WAPD-TM-678, Bettis

Atomic Power Laboratory, January 1967.  ;

9 M.G. Natrella, Experimental Statistics National areau of Standards, Handbook 91 August 1963.

10. J.M. Cano et al., "Supewriticality Through Optimum

!!oc ration in Nuclear Fuel Storage," Nuclear Technolouy,  ;

V:,) B pp. 2 M.-2 60, May 19 8 0.

11. S.L. S4 ner, '

,'diation Tests of Boraflex" Nusurtec Inc @ i. F - 107 (Preliminary), November 1987.

s 4-2.

. c j i

i APPENDIX A BENCHMARK CALCULATIONS B

i l

l A-1 i

~ ,_

1. INTRODUCTION AND

SUMMARY

The objective of this benchmarking study is to verify both the AMPX (NITAWL)-KENO (Ref.1) methodology with the 27-group SCALE cross-section library (Ref. 2) and the CASMO-2E code (Ref. 1 3). for use in criticality calculations of high density spent fuel storage racks. Both calculational methods are based on transport theory and have been benchmarked against critical experiments that simulate typical spent fuel storage rack designs as real-istically as possible. Results of these benchmark calculations with both methodologies are consistent with corresponding calcu- l 1ations reported in the literature and with the requirements of Regu btory Guide 3.41, Rev. 1, May 1977.

Results of these benchmark ca l.cula tions show that the 1

27-group (SCALE) AMPX-KENO calculations consistently underpredict <

the critical eigenvalue by 0.0106 i 0.0048 o k (with a 95% proba-bility at a 95% confidence level) for critical experiments l selected to be representative of realistic spent fuel storage rack configurations and poison worths . Similar calculations by Westinghouse suggest a bias of 0.012 t 0.0023, and the results of ORNL analyses of 54 relatively "claan" critical experiments show a biac of 0.0100 k 0.0013.

Similar calculations with CASMO-2E for clean critical exI. .iments resulted in a bias of 0.0013 0.0018 (95%/95%).

CASMO-2E and AMPX-KENO int 9rcomparison calculations of infinite arrays of poisoned cell configurations show very good agreement

and suggest that a bias of 0.0013 0.0018 is the reasonably expected bias and uncertainty for CASMO-2E calculations .

• Validation of Calculational Methods for Nuclear Criticality safety. (See also ANSI N16.9-1975.)

A-2 Rev. 1 8/12/88

The benchmark calculations reported here indicate that either the 27 group ( S CALE) AMPX-KENO or CASMO-2E calculations are acceptable for criticality analysis of high density spent fuel storage racks. The preferred methodology, howevar, is to perform independent calculations with both code packages and to 1 utilize the higher, more conservative value for the' reference design infinite mult' plication factor.

2. AMPX (NITAWL)-KENO BENCHMARK CALCULATIONS ,

Analysis of a series of Babcock & Wilcox (B&W) critical experiments (Ref. 4), which includet son.e with absorber plates _l 1 typical of a spent fuel rack, is summarized in Table 1 as calcu-lated with AMPX-KENO using the 27-group SCALE cross section library and the Nordheim resonance integral treatment in The mean (and standard deviation of the mean) for these NITAWL. 1 calculations is 0.9894 i 0.0019. With a ono-sided tolerance

] factor (K = 2.502), correspond 3r.g to 95% probability at a 95% ,

confidence level (Ref. 5), the cuculational bias is +0.0106 with l1 t an uncertainty of *0.0048.

Similar calculational deviations reported by Westinghouse (Ref. 6) are also shown in Table 1 and suggest a bias of 0.012 0.0023 (95%/95%). In addition, ORNL (Ref. 7) has analyzed some 54 critical experiments using the same methodology, obtaining a mean bias of 0.0100 0.0013 (95%/95%). These published results 1 are in good agreement with the results obtained in the pres at analysis and lend further credence to the validity of the 27-group AMPX-KENO calculational model for use in criticality analy- j sis of high density spent fuel storage racks. Variance. analysis of the data in Table 1 suggests the possibility that an unkpovn factor may be causing a slightly larger variance than might be  ;

expected from the Monte Carlo statistics alone. However, such a factor, if one truly exists, is too small to be resolved on the

basis of critical-experiment data presently available. No trends  !

l A-3  !

Rev. 1 8/12/88

l Table 1 RESULTS OF 27-GROUP (SCALE) AMPX-KENO CALCULATIONS OF B&W CRITICAL EXPERIMENTS Westinghouse Experiment Calculated Calculated-meas.- l Number. k egg a k,gg l j I 0.9889 0.0049 -0.008 II 1.0040 0.0037 -0.012 l III 0.9985 0.0046 -0.008 l IX 0.9924 0.0046 -0.016 l1 )

X 0.9907 0.0039 -0.008 l XI O.9989 iO.0044 +0.002 XII 0.9932 0.0046 -0.013 XIII 0.9890 0.0054 -0.007 XIV 0.9830 0.0038 -0.013 XV 0.9852 iO.0044 -0.016 XVI 0.9875 0.0042 -0.015 XVII 0.9811 10.0041 -0.015 l XVIII 0.9784 i0.0050 -0.015 XIX 0.9888 0.0033 -0.016 )

l XX 0.9922 0.0048 -0.011 '

XXI O.9783 t0.0039 -0.017 Mean 3.9894 i0.0011(1) -0, 20 0.0010 Bias 0.0106 0.0019(2) I

.; 0.0120 h 0.0010 i Bias (95%/95%) 0.0106 0.0048 0.0120 t 0.0023 Maximum Bias 0.0154 0.0143 l

(1) Calculated from individual standard deviations. 1 (2) Calculated from k egg values. l 1

A-4 Rev. 1 8/12/88

in kegg with intra-assembly water gap, with absorber plate i reactivity worth or with soluble poison concentration, were identified.

3. CASMO-2E BENCHMARK CALCULATIONS 3.1 GENERAL The CASMO-2E code is a multigroup transport theory code utilizing transmission probabilities to accomplish two-dimen-sional calculations of reactivity and depletion for BWR and PWR fuel assemblies. As such, CASMO-2E is well-suited to the criti-cality analysis of spent fuel storage racks, since general practice is to treat the racks as an infinite medium of storage cells, neglecting leakage effects.

CASMO-2E is closely analogous to the EPRI-CPM code (Ref. 9) I and has been extensively benchmarked against hot and cold critical experiments by Studsvik Energiteknik (Ref. 3) . Reported 1 analyses of 26 critical experiments indicate a mean k egg of 1.000 1 0.0037 (la). Yankee Atomic (Ref. 10) has also reported results of extensive benchmark calculations with CASMO-2E. Their ]

analysis of 54 Strawbridge and Barry critical experiments using the repo. ed buckling indicates a mean of 0.9987 Oc0009 (lo) ,

or a bias of 0.0013 ! 0.0018 (with 95% probability at a 95%

confidence level). Calculations were repeated for seven of the  ;

1 Strawbridge and Barry experiments selected at rancom, yielding a I mean kegg of 0.9987 0.0021 (la), thereby confirming 1. hat the cross-section library and analytical methodology being used for l

l Significantly large trends in k gg with water gap and with ab-sorber plate reactivity worth ha,ve been reported (Ref. 8) for AMPX-KENO calculations with the 123-group Gl.A-THERMOS library.

A-5 l

Rev. 1 8/12/88 L s. s .--

the present calculations are the same as those used in the Yankee analyses. Thus, the expected bias for CASMO-2E in the analysis of "clean" critical experiments is 0.0013 1 0.0018 (95%/95%).

3.2 BENCHMARK CALCULATIONS CASMO-2E benchmark calculations have also been made for the B&W series of critical experiments with absorber plates, simu-lating high density spent fuel storage racks. However, CASMO-2E, as an assembly code, cannot directly rapresent an entire core configuration without introducing uncertainty due to reflector constants and the appropriateness of their spectral weighting.

For this reason, the poisoned cell configurations of the central assembly, as calculated by CASMO-2E, were benchmarked rgainst corresponding calculations with the 27-group (SCALE) AMPX-KENO code package. Results of this comparison are shown in Table 2.

Since the dif ferences are well within the normal KENO statistical variation, these calculations confirm the validity of CASMO-2E calculations for the typical high density poisoned spent fuel rack configurations. The differences shown in Table 2 are also consistent with a bias of 0.0013 0.0018, determined in Section 3 .1, as the expected bias and uncertainty of CASMO-2E calcula-tions.

Yankee has attempted such calculations (Ref. 10) using CASMO-2E generated constants in a two-dimensional, four-group PDQ model, obtaining a mean keff of 1.005 for 11, poisoned cases and 1.009 for 5- unpoisoned cases. Thus, Yankee benchmack calculations suggest that CASH 0-2 E tends to slightly overpredict reactivity.

A-6 i .

o .

j Table 2 RESULTS OF CASMO-2E BENCHMARK (INTERCOMPARISON) CALCULATIONS

g. (1)

B&W Experiment No.III AMPX-KENO (2) CASMO-2 E Ak XIX 1.1203 i 0.0032 1.1193 0.0010 XVII 1.1149 0.0039 1.1129 0.0020 XV 1.1059 i 0.0038 1.1052 0.0007 Interpolated (3) 1.1024

• 0.0042 1.1011 0.0013 XIV 1.0983 i 0.0041 1.0979 0.0004 XIII 1.0992 0.0034 1.0979 0.0013 Mean i 0.0038 0.0011 Uncertainty t 0.0006 BWR fuel rack 0.9212 t 0.0027 0.9218 -0.006 (1) Infinite array of cer. tral assemblies of 9-assembly B&W criti-cal configuration (Ref. 4).

(2) k from AMPX-KENO corrected for bias of 0.0106 ak.

(3) Interpolated from Fig. 28 of Reference 4 for soluble boron concentration at critical condition.

l l

l i

l A-7

REFERENCES TO 1.PPENDIX A

1. Green, Lucious, Petrie, Ford, White, Wright, "PSR-63/AMPX-1 l 1 (code packace), AMPX Modular Code System for Generating Coupled Mulcigroup Neutron-Gamma Libraries from ENDF/B,"

ORNL-TM-3706, Oak Ridge National Laboratory, March 1976.

L. M. Petrie and N. F. Cross, "KENO-IV, An Improved Monte l 1 Carlo Criticality Program," ORNL-4938, Oak Ridge National Laboratory, November 1975.

2. R. M. Westfall et al., "SCALE: A Modular Code System for j 1 Performing Standardized Computer Analyses for Licensing Evaluation," NUREG/CR-0200, 1979.

W. E. Ford, III et al . , "A 218-Neutron Group Masta Cross- l1 section Library for Criticality Safety Studies," ORNL/TM-4, 1976.

3. A. Ahlin, M. Edenius, H. Haggblom, "CASMO - A Fuel Assembly l 1 Eurnup Program," AE-RF-76-4158, Studsvik report.

(proprietary).

A. Ahlin and M. Edenius, " CAS MO -A Fast Transport Theory Depletion Code for LWR Analysis ," ANS Transactions, Vol. 26, l 1

p. 604, 1977.

M. Edenius et al., "CASMO Benchmark Report," Studsvik/RF- l 1 78/6293, Aktiebolaget Atomenergi, March 1978.

" CASMO-2E Nuclear Fuel Assembly Analysis, Application Users l1 l Manual." Rev. A, Control Data Corporation, 1982. '

4. M. N. Baldwin et al., "Critical Experiments Supporting Close l1 Proximity Water Storage of Power Reactor Fuel," BAW-lf84-7, The Babcock & Wilcox Company, July 1979.

l

5. M. G. Natrella, Experimental Statistics, National Bureau of l1 Standards, Handbook 91, August 1963.

6. B. F. Cooney et al., " Comparisons of Experiments and il Calculations for LWR Storage Geometries," Westinghouse NES, ANS Transactions, Vol. 39, p. 531, November 1981.

7. R. M. Westfall and J. R. Knight, "Scale System Crocs-section l1 Validation with Shipping-cask Critical Experiments," ANS Transactions , Vol . 33, p. 368, tiovember 1979.

8. S. E. Turner and M. K. Gurley. "Evaluation of AMPX-KENO 11 eenchrrark Calcula tions for High Density Spent Fuel Storage Racks," Nuclear Science and Engineering, 80(2): 230-237, February 1982.

A-8 Rev. 1 8/12/83

(

e s* e REFERENCES TO ~ APPENDIX A (Continued)  !

9. "The EPR2-CPM Data Library," ARMP Computer Code Manuals, l 1 i 3 Part II, Chapter 4, CCM3, Electric Power Research Institute, '

November 1975.

10. E. E. Pilat, "Methods for the Analysis 'of Boiling Water Reactors (Lattice Physics)," YAEC-1232, Yankee Atomic l1 Electric Co . , De'cember 1980.

11. L. E. Strawbridge and R. F. Barry, "Criticality Calculations l 1 for Uniform, Water-moderated Lattices," NSE 23, 58, September 1965.

P i

.1 1

4 A-9 Rev. 1 8/12/85

. - - _ _ _ _ _ _ _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ - _ _ - - _ _ _ _ _ _ _ . _ - _ _ _ _ - _ _ _ . . _ _ _ - - - _ _ - _ _ - _ - - _ _ _ _ - _ _ _ _ _ _ - _ - -