ML20149K223
| ML20149K223 | |
| Person / Time | |
|---|---|
| Site: | Maine Yankee |
| Issue date: | 01/31/1988 |
| From: | Cacciapouti R, Carpenito F, Napolitano D YANKEE ATOMIC ELECTRIC CO. |
| To: | |
| Shared Package | |
| ML20149K158 | List: |
| References | |
| YAEC-1622, NUDOCS 8802230413 | |
| Download: ML20149K223 (57) | |
Text
{{#Wiki_filter:. .. .
+ .. -
r
,a ..- YANKEE ATOMIC ELECTRIC COMPANY 3 m
.a
, N
!l>
4
~;'
[' ; validation of the YAEC Criticality Safety Methodology By -
.7 D. G. Napolitano F. L. Carpenito '
N.7l, January 1988
.1- .4 Y
a
- ,9.
e - lU l 1. 0 NOTICE - w THE ATTACHED FILES ARE OFFICIAL RECORDS OF THE
?;
DIVISION OF DOCUMENT CONTROL. THEY HAVE SEEN co . , . CHARGEO TO YOU FOR A LIMITED TIME PERIOD AND l BRANCHBE MUST 016.RETURNED TO THE RECORDS FACILITY Mt%E 00 NOT SEND DOCUMENTS , l: CHARGED OUT THRE!GH THE MAIL. REMOVAL OF ANY
- b. ' PAGE(S) FROM DOCUMENT FOR REPRODUCTION MUST BE REFERRED TO FILE PERSONNEL.
p; .- DEADLINE RETURN DATE H i L, - i 1 - RECORDS FACILITY BR ANCH F l o- ' L EL-uwaun e r.m wt U n %n.,rt(e?a,.
- - @I nV
.. - ~ Docket # Co-to#}
8802230413WO$fL-vev1 Canuvi
- TTCW.* C8tC l-P,DR ADOCK 0500030j -
Dat*0 A D W af NG '
4 t
<- n
- ? U l
O ' j? 5 a Validation of the YAEC f ri Criticality safety Methodology
-*' By f;
D. G. Napolitano F. L. Carpenito r~ g, January 1988 L. 1: c_ Prepared by: aw 4/M wo /![/88
') D. G. Napolitah% N~ucleaf Engineer / 'Da te
-[i. Reactor Physics Group
- p. Nuclear Engineering Department I. ..
Prepared by: Y u b O!Ob Ca ryenito,7nginee r Date ' .' F. L. ! ! [d, Reactor Physics Grouy Nuclear Engineering Department Approved by: /![
/ Wate M
R gJ . Cacc outi, Manager l , _. Reactor P ics Group I J. , Nuclear E intering Department lC Approved by: //[!IO Date B. C. Slig)r, Director , c, Nuclear E6gineering Department Yankee Atomic Electric Company i, Nuclear Services Division 1671 Worcester Road Framingham, F.assachusetts 01701 n , e b.- FJJUJ0i1Y DSCXET RLE COPY
DISCLAIMER OF RESPONSIBILITY
-r This document was prepared by Yankee Atomic Electric Company l
("Yankee"). The use of information contained in this document by
-/ anyone other than Yankee, or the Organisation for which this document was prepared under contract, is not authorized and, with
(" respect to any unauthorized use, neither Yankee or its officers, directors, agents, or employees assume any obligation, p responsibility, or liability or make any warranty or U representation as to the accuracy or completeness of the material contained in this document. C y T' 4 {. 7.. kw I 1h l
- f
%' ' s
.(. b 4 'Yp a.. l l-4
. . i f
AB8 TRACT l' , The YAEC criticality safety methodology has three n, calculational paths: NITAWL-S/ KENO-Va Monte Carlo analysis, CASMO-3 l} integral transport analysis and CA8MO-3/CKART-2/PDQ-7 diffusion theory analysis. Twenty one Babcock & Wilcox (B&W) fuel storage rack critical axperiments are modelled with this methodology. This
- " investigation was done to validate the methodology for licensing
- { calculations of fuel storage criticality. A methodology bias and 95/95 uncertainty is determined for each path, and the application f of the bias and uncertainty to fuel storage criticality analysis is
- i. Illustrated.
o I' e 'l. P. i r l L. l(
- l?
I-i, IE
*a U. ;
.r. l- , a)
- s
] 'l I
- iii -
J h
~
r' i
. , i j,' .
c' ACKNOWLEDGMENTS 4. j:. ) The authors would like to express their gratitude to Peter Rashid and Malte Edenius of Studsvik of America for their contribution to the CASMO-3 benchmarking effort in this report. i 4 ( *.*
^
5
\
C' h *_ - f. f- , t .; . A v
?.
l' - v. f-
- u. .
I' L:
- iv - ; ?
,;. . - i
.5 f
.i
'- TABLE OF CONTENTS y.
Page
.f 11 l
- t. DISCLAIMER ........................................... i
-, y ,
111 l ABSTRACT ............................................. r~ iv
, ACKNOWLEDGMENTS ......................................
3 0. V TABLE OF CONTENTS ....................................
' r :.
Vi (, LIST OF FIGURES ...................................... vii
- r. LIST OF TABLES .......................................
.t .
.t' 1 l.0 INTRODUCTION .....................................
.............. 4 2.0 DESCRIPTIO!! OF CRITICAL EXPERIMENTS h'
- 4 ............ 22 3.0 NITAWL-C/ KENO-Va CRITICALITY ANALYSIS r 3.1 NITAWL/ KENO Modelling ....................... 22
}J 4 3.2 Results and Statistical Analysis ............ 23 26 3.3 Trends ......................................
T:' 33
- l , 4.0 CASMO-3 CRITIC'ALITY ANALYSIS .....................
6 4.1 CASMO Lattice and Reflector Modelling ....... 33 34 p 4.2 CASMO/CMART/PDQ Modelling ................... 36 0 4.3 Results and Statistical Analysis ............ 37 l 4.4 Trends ...................................... Ib 5.0 APPLICATION OF UNCERTAINTY TO SPENT FUEL RACK 42 CRITICALITY ANALYSIS ............................. 5.1 Criticality Safety Design Criteria .......... 42
-' 5.2 NITAWL-S/ KENO-Va Uncertainty Calculation .... 43 5.3 CASMO-3 Uncertainty calculation ............. 44 1r z g- 46 b-
6.0 CONCLUSION
S ...................................... i 48
!2 REFERENCES .......................................
- j. .
- 4. l
. t I I h.
I 1.
- b.
- i -V - ,
.. t
(
---- --~. _ _ . . _ _ _ _ _ , _ _ _ _ _ _ _ _ _ _ _ _ _ __ _
t t [. LIST OP PIGURES V' Mumber _ Tltie Paqe
!' 1.1 7AEC Criticality Eafety Methodology 3
[5 2.1 Core I Loading Diagram - Cylindrical Base Case 9 2.2 Core II Loading Diagram - Nine Arrays with Zero 7 Pin Pitch Separation 10 l' 2.3 Core III Loading Diagram - Nine Arrays separated by One Pin Pitch 11 [ 2.4 Core IV Loadino Diagram - Nine Arrays Separated by One Pin Pitc3 and 84 8 4 C Fins 12
- k 2.5 Core V Loading Diagram - Nine Arrays Separated by Two Pin Pitches and 64 5 4 C Pins 13 j
'I 2.6 Core VI Loading Diagram - Nine Arrays Separated
! i .- by Two Pin Pitches and 64 8 4 C Pins 14 2.7 Core VII Loading Diagram - Nine Arrays Separated
((. bf Three Pin Pitches and 34 B 4C Pins 15 _. 2.8 Core VIII Loading Diagram - Nine Arrays separated by Three Pin Pitches and 34 B 4 C Pins 16 L 2.9 Core IX Loading Diagram - Nine Arrays Separated {n- Four Pin Pitches 17 ; 2.10 Core X Loading Diagram - Nine Unit Assemblies separated by Three Pin Pitches 18 7-lb 2.11 Cores XI, XIII, XIV, XV, XVII and XIX - Nine Unit Assemblies separated by One Pin Pitch and Metal p "' Isolation sheets 19 2.12 Cores XII, XVI, XVIII and XX - Nine Unit Assemblies i o separated by Two Pin Pitches and Metal Isolation P . sheets 20 i l.
- j. 2.13 Core XXI - Nine Unit Assemblies Separated by Three
!E Pin Pitches and Metal Isolation sheets 21
- L.
3.1 KENO-Va Model of Critical Experiments 29 ii ' f_ l' i f a. 's
- vi -
s. e f.. s LIST OF TkBLES l I, Number Title Page T 2.1 B&W Fuel Storage Criticals - Core Conditions 6 2.2 ruel Pin Properties 7 2.3 B C Fin Properties 7 4 2.4 stainless steel Isolation sheet Properties 8 ,6 4 h' 2.5 Borated Aluminum Isolation Sheet Properties 8 r' ' ,I_ 3.1 KENO-Va Core K,gg Results 30 ,
- - 3.2 KENO-Va Results vs. Gap Spacing 31 a
3.3 KENO-Va Results vs. B-Al Sheet Boron Loading 33 P 4.1 CASMO-3 Lattice K-Infinity Results 38 1, 4.2 CASMO-3/PDO-7 Core K,gg Results 39 'l CASMO-3/PDQ-7 Results vs. Gap Spacing 40 4.3 4.4 CASMO-3/PDQ-7 Results vs. B-Al Sheet Boron Loading 41 A> < y .(..
- ~ >
kll
't . s (1 .
,g i 4. 6 f s I ^"
- vii -
1 L i
. ~ , .. ,
t t Q v-
1.0 INTRODUCTION
<~.
lb The YAEC criticality safety methodology is shown in rigure , [ 1.1. Input to the methodology begins with fuel assembly and storage rack data. This data is transformed into material I k! compositions and geometry input for the methods that follow. The l 0, methodology has three calculational paths: MITAWL-8/ KENO-Va, CASMO-3 and CASMO-3/CRART-2/FDQ-7. II' } The reference calculational method is NITAWL-5/ KENO-Va . The NITAWL-S code prepares a working nuclide library and performs r 238 . Either 123 group or 27 group l' resonance self-shielding for U data can be used in this pathway. The working nuclide library g., (., along with case specific compositions and rack geometry data are . E input to KENO-Va. KENO-Va is a multi-group, Monte Carlo L KENO-Va can analyzed one, two or three criticality code. p dimensional arrays. A variety of boundary conditions and/or (, reflector regions can be modelled. The results from. KENO-Va b analysis are ny gg vs. generation, fluxes and reaction rates.
'f sin A Monte Carlo is stochastic in nature, results will always h "~' have some uncertainty (+ or - a standard deviation).
C' . The auxiliary calculational method is CASMO-3 and/or CASMO-3/ CHART-2/PDQ-7 I '4'bl. CASMO-3 is an integral transport lattice code with a hierarchy of energy condensation and spatial detail r leading to a seven-group, transmission probability model of the
' unit cell, COXY. CASMO-3 is flexible enough to handle up to a ! 19x19 fuel assembly array with storage canister regions, poison sheets and water gaps. If the rack unit cell is simple enough,
. - _ - _,, 4
\: j l
l W In CASMO-3 can be used by itself to study unit cell sensitivity. g, addition, for simple lattices and canister arrays, colorsets
' [.'i (2 by 2 arrays) may be possible. Either 70 or 40 group nuclear
; data can be used in CA8MO-3. The result of CASMo-3 calculation is
" a deterministic K,for a unit cell with no axial or radial h(, leakage. l CAsMo-3 can produce few group cross sections which preserve 6 These cross sections can be processed by unit cell reactivity.
~
~ the CRART-2 code for use in PDQ diffusion theory analysis. Using c
'- PDQ, a refined unit cell analysis or large array studies can be C performed. Also, axial and radial leakage effects can be h
evaluated. I'? Additionally, CAsMo-3 can perform burnup credit analysis. Hot w full power lattice depletions can be executed, and cold sero power i g t; restarts in rack geometry can be performed. The effects of incore i g:, depletion and long term out-of-core nuclide decay can be studied.
- m
However, validation of burnup credit is beyond the scope of the r*. .
- 17 present report.
Le
- The calculational p:ths of the YAEC criticality safety
, s 7)j. ,
methodology will be validated by comparison to 21 B&W fuel storage l critical experiments (6) . These experiments will test the hbility I: of the methodology to predict critical (K,gg = 1). Statistical j! analysis of the 21 calculated K,gg(s) will give a methnd bias and j"" 95/95 methodology uncertainty for each calculational path. 1? I.. 30 l i N, j t l-1. i.,
, ----..,,-,-,,--,,------c . _ , . , . - - , - - - . . . _ , _ - - , - - - . - - , - - - -,- -
p, . v . C
'[' '
Figure 1.1 YAEC Criticality Safety Methodology
, r'
' jk .
FUEL ASSEMBLY DESIGN STORAGE RACK DESIGN
. ') . r^ ;/'~ MATERIAL COMPOSITIONS AND GEOMETRY
[s
.' ,%, s REFERENCE-. - AUXILARY E> CALCULATIONAL METHOD CALCULATIONAL METHOD
\
r 's ja CASMO-3 NITAWL-S s
- INTEGRAL TRANSPORT fr. .
- 3. 1
, ~ , ,
i. r- WORKING UNIT CELL ! l['j NUCLIDE SENSITIVITIES
.- 'g.h,IBRARY
{. \ ' CHART-2 ~ 5.e FEW GROUP X-S
- r; KENO-Va MONTE CARLO I " , . . -
lf .. I Foo-7 x i .' D'iPTUSION THEORY (,, x 6 . - , R El'E'R'EN C"e,; - ' F.- Y,L'4? T l;, ,-- LARGE ARRAY ~3TUDIES l - __.- - . .u. 3, \. ; - > t i "g 4 ' \' _ , s_ , .
~>
b y .- , 3 , i'
\
1 .,
. s N
^
, y \
l \
\
s ' '
s
*-e
- s ., , se a- ;
. s
'n l
d _, _ i ! _ _ __ _ _ _ . - - - -
~ . _ .
I r
2.0 DESCRIPTION
OF CRITICAL EXPERIMENTS l 1 tr ! lf A series of 21 critical experiments were performed by B&W in support of spent fuel storage of LWR type fuel (6) These {, experiments were performed at the CX-10 critical facil,ity. CX-10 qz
- ( is a tank-type facility licensed for the performance of critical experiments with water moderated U02 and mixed oxide fuel H lattices. Each critical experiment, except for Core I, was a 3 X
^ 3 array of fuel assemblies. Each assembly comprised a 14 X 14 3 ' array of fuel rods with an active fuel height of approximately 5 h ft. The pin pitch in all cores was 1.636 cm. The fuel rods 235 contained 2.46 w/o U UO2 pellets which were clad with aluminum. : [. Criticality was maintained in the cores by a combination of water i r height, soluble boron, fixed poison (5 C 4 pins and boral sheets) l b, and array spacing. A summary of the core conditions is given in [ Table 2.1. The fuel and fixed poison properties are summarised in L Tables 2.2, 2.3, 2.4 and 2.5. At a lattice pitch of 1.636 cm, the water-to-fuel ratio of
._ j g 1.88 is comparable to that of LWR fuel assemblies, 1.80 to 1.95.
L. . Therefore, except for the use of aluminum cladding rather than r , zircaloy, the B&W series of critical experiments is reasonably ; representative of LWR spent fuel storage arrays. The experiments comprised three distinct core geometries. , f.
- 1. Core I was a cylindrical array of 438 fuel pins constructed as i
L a base case, see rigure 2.1
- 2. Cores II through IX were a 3 X 3 array of fuel pin clusters i i
?- with each cluster a 14 X 14 array of fuel pins. The fuel l-1.. [' . 1-____--- - _ _ _ - - _ _ __ ___/
.:.. v , .: . .. v. -
. .n t
} clusters were spaced various distances apart and vertically
, aligned by continuous upper and lower grids. In some cases, 3 4 c pins were inserted in the space between the fuel clusters, r see Figures 2.2 through 2.9.
\
- 3. Cores X through XXI were a 3 x 3 array of fuel pin clusters, with each cluster comprising a 14 x 14 array of fuel pins.
However, the four corner positions of each fuel assembly were occupied by threaded aluminum rods. These corner rods positioned the three grid plates and provided the unit assembly framework. The free standing feature of the unit assemblies was necessary to accommodate metal isolation sheets. The isolation sheets were either 0.462 cm thick stainless steel or L 0.645 cm thick borated aluminum. Like Cores II through IX, the fuel assemblies were spaced various distances apart, see Figures 2.10 through 2.13. b k l 1 6 % k ( i
~
l
-s-c .. . . -- -
l [ '
-* rigure 2.1 Core I Loading Diagram - cylindrical Base case P
i f ...... ;
$,? ..
i t
' t. '
.l . f' F. , [' Fuel Rod Position !
. t } ', ;
Pl s. l d' b-c I ! f,: i 1)' % r. h . T, i S-L i e f f 1 9 h I~ 1 l 1
-. - ~_- _ _ _ -_. __.__ ___,._ _ ___ _ ____ _ _ _ _ __ _ _ _ _ , _ _ _ _ _ _ _ _ _ _ _ _ _ _
_ _i
i
't i: rigure 2.2 Core II Loading Diagram - Mine Arrays e ,-
with zero Fin Fitch separation g ' te . f
.l!
=r 'l, .l-t y .., l i e- '[ .- f-t. O Fuel Rod Position 0, l' r
,1 .
t..
. . . i i
L
. l
i Pigure 2.3 Core III Loading Diagram - Nine Arrays
.O' separated by One Pin Pitch Y.
t-i n l' c t
*, s i
- t. .
r-e O Fuel Rod Position (b n 1-L + i. be -
=
4
, - y -~~ - - - ---
w-w ~~wv*+-, .ev -.--w-----~~r- -----r-m - - - - - - - - - -
4 t 1 Figure 2.4 Core IV Loading Diagram - Mine Arrays F Separated by One Pin Pitch and 84 34C Fins
- b. .
- i. 'N )
s = p n a a m
s a e a t-e s m a m =
a E f._ m n m a a m (I. .
................ m . . . ..
m u a m h,
' s a \. . m a e u o- m a
e . 1- - o ! ',. O Fuel Rod Position U. ., e B C Pin S .g. . l, E I r I o s t 1
y . - r 6 r. Pigure 2.5 core v Loading Diagram - Nine Arrays c Separated by Two Pin Fitches and 64 34C Fins (- . a m
. f, r, e o e n a .
' a a aa aaa a a a a eaa aa aa I
e n
,fi
. e e n , .
If
- . e e i w w w w w w 4 w w w w # r w w w ww
= -
l- e e i u e e 'I-e n e s l hip' O Fuel Rod Position
- 3 4C Fin w.
p p s. 1 I[ l l he J;
- 13 - .
L ig* lI
3 i i
- f. .
P Figure 2.7 Core VII Loading Diagram - Mine Arrays Separated by Three Pin Pitches and 34 34C Fins
\.
I g. f. l l*. e e 1
-' \
,. e e i
e e r 1, a a a . . . . .
\
_j e n (.. t e o e n
.- I w w w w , w , w e e ,
<:, 1 1
- k :.- . ,
r-1 [ . ', D Fuel Rod Position
- 3 C Pin 4
.'r '
b* he ' t
~
m 1 V
. ^ -
( . . . ;l r-(, 4 Figure 2.8 Core VIII Loading Diagram - Nine Arrays [' . separated by Three Pin Pitches and 34 34C Fins l..
. I.
ic e e f'. I I -. e e r- , , i . e e e e e e e *
. e f .'
L
. . e f- e
- l. ' e ,
e e o e e e e e e
- f. ,
L. , F e . '
- L n
I L s- . O Fuel Rod Position e 54C Pin i Y.
- 4%
- y. -
i *
.s ;
I c- , 6 4 l... > 1 i 4
- 16 - i
}
l' i I l I
. l
, 7 L
. i
(.- '
' Figure 2.9 Core IX Loading Diagram - Mine Arrays separate by Four Fin Fitches 1: ,[.
I', t 4.l '. (-
. l. ,. .c ,i.
- l. .
n p . t.
}s.
- 1. ,
[L
~
(C D Tuel Rod Position f '. .,'. l
)*_' g P.1 1 .
s. I _ l 1, i p . 4 g l w - -, ~. _ , - , ,_, , _ . . , _ - _ . _ , ,_,. , _ _ _ _ , , - _ _ . _ _ _ , . , , _ , , . - , _ , ,
7_ . t r ('. Figure 2.10 Core X Loading Diagram - Bline Unit Assemblies
- Separated by Three Pin Pitches m:
4 c:- i. 1 F L
- r. .
V I.. 7.. t-4 If e O Tuel Rod Position
- Threaded Aluminum Rod i:'
6-
-ls. .
t. w 6 'e em e. 9
4 Figure 2.11 l Cores XI, XIII, XIV, IV, XVII and XIX - Nine Unit Asseeablies l Separated by one Fin Fitch and Metal Isolation Sheets l I l , .t'
- l,i
- c !
I' l o i r-t Mk l b
- r. . .
. ( .. j 1; r t-f t l E
... j L3 D Fuel Rod Position e Threaded Aluminum Rod I b: Isolation Sheet L. -
I a s_, i i f i
'i !
i I
r f- ' rigure 2.12 cores XII, XVI, XVIII and XX - Nine Unit Assemblies p separated by Two Pin Pitches and Metal Isolation Sheets - t t-9 { u e, i fi. 1. e-
; h t
I' f' t t_ r, e-i p Li-Q . i o Fuel Rod Position {, e Threaded Aluminum Rod i Isolation Sheet ) l i 1 l i i
't l
\
l i l I 1
- +
1 t i r i rigure 2.13 Core XXI - Mine Unit Assemblies lr.
't separated by Three Fin Fitches and Metal Isolation sheets
- r; ,
'L , ;I+ -[ . ,.
t . :. . >. F I' '
.b
'e t
'. J.
. J.
f -- .f M g. s lv.T. Fi c; - O Fuel rod Position 3
- Threaded Aluminum Rod !
- Isolatien Sheet ,
i h t
. .. n . . .
l-r
, 3.0 NITAWL-5/ KENO-Va CRITICALITY ANALYSIS i
~
3.1 NITAWL/ KENO Nodelling ]}L4 Working nuclide library creation and resonance self shielding was done with the NITAWL-5 code. Working libraries r for both the 123 and 27 group analysis were created. Resonance (}
, self-shielding was accounted for in the U 238 isotope only. This i
v was handled differently for interior, edge and corner fuel pins of each assembly by the use of different Dancoff factors. These were: .1925, .1346 and .0868 for interior, edge and corner fuel pins, respectively(6) . This produced three different U 238 data L-sets for the criticality analysis, n h The neutron multiplication factor, K,gg, of each benchmark core was calculated using the Monte Carlo criticality code i l-KENO-Va. The portions of the core that were mocked up by the ! [] L basic KENO-Va model are shown in Figure 3.1. The computer l representation explicitly modelled each fuel rod, each 3 4 C pin, l the isolation sheets, the moderator and core tank in the X-Y direction. In the I direction, the cores were modelled up to j fn. ! L, the critical water height. Reference 6 and others indicate ; t
- that the fuel and structure above the critical moderator height i does not contribute discernibly to neutron multiplication, j Therefore, the fuel and support structure above the critical water height were not modelled. Based on the adjusted {
measurements, the critical moderator height was set to 145 cm for Cores I through IX and to 150 cm for Cores X through XXI. i
. - - _ - -- L
. . - ~_ -
l' 4 I C 3.2 REsULTS AND STATISTICAL ANALYSIS r The major objective of the KENO-Va benchmarking of the B&W '[ ' criticals is to quantify the method bias, AKb, and method ; uncertainty, e,, in calculating K,gg. The method bias is the average difference between the each measured K,gg and each
- r
]*; calculated K,gg. The 'athod uncertainty is the standard
, deviation of the distriaution of calculated K,gg(s). The ;
- , statistical analysis of KENO-Va results for the latter quantity J is complicated by the fact that Monte Carlo calculations C
introduced a random component into the distribution of . ] calculated K,gg(s). Reference 7 gives a method for calculating the systematic U and random effects of Monte Carlo analysis, systematic effects j
- are those associated with the calculational methodology, i.e. [
the code and data sets. Such effects are characterised by the [' - method bias, and the method variance (square of the standard deviation). In contrast, random effects are those that are stochastic in nature, i.e. Monte Carlo error and mechanical ; g tolerances. Randon Monte Carlo error is characterised by an b- average Monte Carlo variance. This random component must be subtracted off of the variance calculated from a sample of Monte ! Carlo calculations in order to obtain an unbiased estimate of , the method variance. Reference 7 recommends the following equations for i calculations of the systematic and random effects in Monte Carlo validation studies: [ _ _ _ _ _ _ _ , . . _ , n._- , - . _ - - -
- g. .
I e 2 IKg/eg i Average K,gg s Y = 1 (1)
}. n E(1/e )
- 1 n
~~
-2 Ze *NG g Average Monte Carlos e - 1 (2)
Variance n ENG g
,F -
1 n Method Blas AK b
" 1*Z(K"i - Kg) (3) 7 n1
- l .-
l,f 2 n*E(Kg -E) /e -2
- Method variances e, = 1 - e (4) n :
(n-1)*E(1/e ) f-
, 1 ,
2 where: . Ka g = Measured K,gg of Critical Experiment i Kg = Monte Carlo K,gg of Critical Experiment i (- 2 e = Variance of Monte Carlo Calculation for critical Experiment i
.* NG g = NumP,r of Neutron Generations used in the Monte car.Ao Calculation of Critical Experiment i n = Number of Critical Experiments calculated by Monte Carlo
. Note, that in the calculation of R, Equation 1 weights each K g by 1/e . This gives higher weight to calculations which t., have higher accuracy, i.e. lower variance. In the calculation of average Monte Carlo variance, e 2, equation 2 weights each Monte
' Carlo variance by the number of generations executed. This gives i higher weight to a variance produced from a greater number of L ,
generations. In the calculation of the variance of the K g(s), equation 4 weights each square error, (Kg -R) , by 1/eg. Again,
=. - _ _ _ _ . .- ._
this gives higher weight to calculations which have higher accuracy. Also, in equation 4, ;2 (the random component) is subtracted from the calculation of the variance of the p L distribution of calculated K,gg(s). This produces an unbiased .g, estimate of the method variance, e 2, I' KENO-Va Monte Carlo calculations for each core representation I were performed with 30,000 neutron histories in combinations of ( 600 neutrons per generation for 50 generations or 300 neutrons per r: [ generation for 100 generations. This resulted in a average Monte [_, ,. Carlo error, e, of 1 00364 for the 123 group data and 1 00342 for the 27 group data. The Monte Carlo error was kept to a minimum by skipping or deleting the early generations and selecting the value of K,gg with minimum standard deviation. The K,gg(s) for the 21 cores are shown in Table 3.1. Column 1 shows the measured Kggg(s) and columns 2 and 3 are the KENO-Va L_ [, results for the 123 and 27 group data, respectively. The
, statistics, E e, for the measurements is 1.0006 2 0011. The average K,gg, E, for KENO-Va with the 123 group library is 0.99686 T with a method uncertainty, e,, of 1 00525. The bias, AK b' If this set of calculations is .00395. The average K,gg, E, for KENO-Va with the 27 group library is .99051 with a method ]
l uncertainty, e,, of 1 00487. The bias, AK b, for this set of U calculations is .01059. U- It is worth noting that the method uncertainty for both i ' libraries is about the same, but the bias for the 27 group library is about 1% AK. However, the 27 group library cases executed about 20% faster on average, providing about a 20% cost savings.
m . p.
- a. .
T . I NRC requires that the methodology uncertainty be quantified . [, at a 95/95 one sided tolerance level (8) This means that K,gg I will be below the computed value with 95% probability at a 95% confidence level. The 95/95 tolerance factors are tabulated in Reference 9. These factors are multiplied by the method r. y
- i. uncertainty calculated above. For the set of 21 benchmark p calculations computed, this is a factor of 2.371. This results in a 95/95 methodology uncertainty of .01245 and .01155 for KENO-Va i l
with the 123 group and 27 group data, respectively. IM u, 3.3 TRENDS 1 The B&W critical experiments were performed with different { water gaps between fuel assemblies, boral boron loadings, number of 5 4 C pins, soluble boron level and to a lesser extent system temperature. When the K,gg(s) are collected and analysed two major trends are observed in the K,gg(s) using the 123 group data, a trend with gap spacing and a trend with borated aluminum boron loading. However, no significant trends are observed in the K,gg(s) using the 27 group data, see Tables 3.2 and 3.3. As one can see from Table 3.2, there is an negative trend of
.K,gg vs. increasing gap spacing when using the 123 group data.
This leads to a significant underprediction (-1%AK) of K,gg for j, Core IX. A linear least squares fit of the data indicates the
- r. slope of this trend is -0.0021 AK per cm. From Table 3.3, a ppsitive trend of K,gg with borated aluminum boron loading is seen. A linear least squares fit of this set of data indicates the slope of the trend is +0.0046 AK per w/o soron.
L.
7he trends exhibited by KENO when using the 123 group data were also noticed by Eng I I, and S. Turner and M. Gurley(10) , but i [ l ( they did not analyse the criticals using the 27 group data. Don ! Marris of RFI has given us pr.rtial explanation of the observed trends and biases IIII. Note, the 123 group library is an old f adjusted data set based on ENDF/II data and the 27 group data is !
- based on modern ENDF/IV data. Marris explained the trend with gap spacing for the 123 group data due to a discrepancy in O(n,s) l I' I data. This discrepancy leads to greater absorption in the water gap. Modern data sets such as the 27 group data are corrected for l j; this. Thus, there is no trend with gap spacing for the 27 group - data. The observed bias of 1% AK for the 27 group data is due l 238 absorption. This b, mostly to the overprediction of U f II8 absorption is observed consistently in r= overprediction of U l l
calculations of K,gg with cross sections based on modern ENDP { I[ data (12) . In fact, the data libraries for CAsMo-3 were recently L adjusted for this and other effects. Originally, the 123 group 0 g data set was adjusted based on Hellstrand measurements of U f l
- resonance absorption. The observed trend with boral boron using the 123 group data is not well explained, and is based on a l l
. limited set of observations which may not be fully indicative of a l l
trend. I l Theoretically, the trends indicated using the 123 group data should be applied to the bias for a particular configuration of gap spacing and boron loading. However, the applicability of these trends is questionable with respect to high density fuel
~
racks that YAEC analyzes. Such designs have flux trap gaps and
- 27
.s e
(* I 1. i boral loading in the range of 20 to 40 w/o 3. Therefore, at the
~
Present time, these trends will not be applied in the uncertainty methodology, f .l C-I f 6 N we + I h 4 L b - N*
.ws 8
e w
\
l
- 2e - !
I i r i s
+ FiTure 3.1 RENO-Va Model of Critical Esperiments AREA AB0VE WATER SATETY CONTROL V01D IN MODEL -%
RCD BLADES
- j. . . . en:w...w m .mr_~ __
i AL. TOP GRID -
.Q]{@@
- WATER I l I I
[ AL BOTTOM GRID -
% u 4. I II 7,
g- 1 1- 2 i + AL, BASE PLATE
]
N AL, CORE TANK
~
NOT MODELED C MODELED L: a p- , IL*. r: b.$ , r;. . i-l I. l s u c.
- 29 -
I
L Table 3.1 KENO-Va Core Keff Results Core Measured
- 123 Group 27 Group
+ Kg NG g Kg i og NG g i Kag eg i og
- I 1.0002 1 0005 .99815 1 00382 50 .99607 1 00398 49 g II 1.0001 1 0005 1.00085 g .00294 50 .99803 1 00300 50 f III 1.0000 g .0006 .99886 3 00336 44 .99433 1 00337 50 IV .9999 1 0006 1.00107 g .00416 48 .98432 1 00385 49 V 1.0000 1 0007 1.00093 1 00441 50 .99719 1 00288 50 VI 1.0097 1 0012 1.00188 1 00364 44 .99567 1 00377 50 P VII .9998 1 0009 .99124 1 00342 49 .99558 1 00329 45 b VIII 1.0083 i .0012 .99937 1 00381 50 .99873 1 00333 50 r..- IX 1.0030 1 0009 .99116 1 00365 47 .99092 2 00355 50 L' . X 1.0001 1 0009 .98851 1 00362 50 .99141 1 00393 49 XI 1.0000 g .0006 1.00673 g .00360 50 .99606 1 00292 50 XII 1.0000 1 0007 .99105 1 00402 50 .98554 1 00320 49 XIII 1.0000 g .0012 1.01049 1 00402 50 .98764 g .00306 48 F XIV 1.0001 1 0012 1.00544 1 00367 50 .98629 1 00342 50 C .99344 1 00344 47 .98766 1 00306 50 XV .9998 1 0016 y XVI 1.0001 1 0019 .98860 1 00348 100 .98432 1 00346 60
'... XVII 1.0000 1 0010 1.00272 1 00314 46 .98204 1 00312 50 XVIII 1.0002 1 0011 .98955 1 00317 96 .98659 1 00344 100
- i. XIX 1.0002 1 0010 .99591 1 00276 48 .99065 1 00370 49 XX 1.0003 1 0011 .99396 1 00460 50 .99176 1 00364 50 h 'XXI .9997 1 0015 .98856 1 00368 50 .97829 1 00325 45 m-R+~ -
1.0006 + .0011
.99686 + .00364
.99051 + .00342
(,
- e, 1 00525 g .00487
.01245 .01155
{ ,' 'm95/95
&E .00395 .01059 b
E ma'
*Trom Reference 6 i s-. 1 i
L i L. 4 l i i j
( t. I-I ' Table 3.2 Ramo-Va mesults vs. Gap Spacing
- r I' Spacing Core 123 Group 27 Group p (ca) i Kg i og NG g Kg i og NG g h 0.0 II 1.00085 1 00294 50 .99803 1 00300 50 1.636 III .99886 + .00336 44 .99433 + .00337 50 IV 1.00107 1 00416 48 .98432 1 00385 49
" XI 1.00673 1 00360 50 .99606 1 00292 50 12 XIII 1.01049 i .00402 50 .98764 1 00306 48 p- XIV 1.00544 1 00367 50 .98629 1 00342 50 !.1 XV .99344 1 00344 47 .98766 i .00306 50 [
XVII 1.00272 1 00314 46 .98204 i .00312 50 ; ]' XIX .99591 1 002/6 48 .99065 1 00370 49 l E+7 - 1.00106 + .00355
.98882 + .00333 l
[: 3.272 V 1.00093 1 00441 50 .99719 i .00284 50 l r,P VI 1.00148 1 00364 44 .99567 1 00377 50 , L XII .99105 + .00402 50 .98554 1 00320 49 l r. XVI .98860 1 00348 100 .98432 1 00346 60 i h, XVIII .98955 1 00317 96 .98659 1 00344 100 l XX .99396 1 00460 50 .99176 i .00364 50 h' i17 .99371 1 00379 .99038 1 00342 f 4.908 VII .99124 1 00342 49 .99558 i .00329 45
-- VIII .99937 1 00381 50 .99873 i .00333 50 g X .98851 + .00362 50 .99141 i .00393 49 L XXI .98856 1 00368 50 .97829 1 00325 45 r-g I+7 -
.99175 + .00364
- .99083._+ .00347 6.544 IX .99116 1 00365 47 .99092 1 00355 50 l.
u e d 4 e
L
- t. ;
- L 7.
1 Table 3.3 KENO-Va Resulta vs. B-Al Sheet Doron Loading , r Boron Core 123 Group 27 Group r (w/c) i Kg i og NG g Kg i eg NG g . f 0.1 XIX .99591 1 00276 48 .99065 1 00370 49 E 00460 50 00364 50 XX .99396 1 .99176 1 C. XXI .98856 1 00368 50 .97829 1 00325 45 i' i I17 .99339 1 00377 .98618 1 00354 0.242 XVII 1.00272 + .00314 46 .98204 + .00312 50 KVIII .98955 + .00317 96 .98659 + .00344 100 r
\f Y17 .99620 1 00316 .98409 1 00334 0.401 XV .99344 1 00344 47 .98766 1 00306 50 j7 XVI .98860 1 00348 100 .98432 1 00346 60 Y17 .99105 1 00347 .98619 1 00328 }'.:- 7 1.257 KIV 1.00544 + .00367 50 .98629 + .00342 50 :
ID i 1.620 XIII 1.01049 1 00402 50 .98764 g .00306 48 l ll. I c !
]c ;
f?' i I s.. '
~
i w. 1 u t I
~
bW D L l
l f- [ 4.0 CABMO-3 CRITICALITY ANALYSIS 4.1 CAsMO-3 Lattice and Reflect _or Modelling I' since the CAsMo-3 lattice geometry is limited, the B&W h criticals are modelled with CASMO-3 generated cross section data and an explicit PDQ model. Two types of CAsMO(s) are executedt a lattice calculation and a reflector calculation. The lattice h. P calculation provides cross sections for the fuel, gap and poison regions. The reflector calculation provides cross sections for the water reflector surrounding the core array. { since all cores except I consisted of 3 x 3 bundles each of 14 x 14 pins, lattice calculations modelled the basic 14 x 14 r' bundle unit. Lattice dreta was generated as a function of array {
,- gap spacing, soluble boron, and poison presence. The poison sheet I
- f. was modelled in the CA5MO-3 as a PWR lattice calculation with a (i channel wall placed in the center of the water gap. The explicit PDQ calculation then places the sheet in the correct location relative to the bundles. The cores with 5 4 C pins in the water gap (IV-VIII) were modelled by a type of colorset calculotten.
O[. , However, this lattice model breaks down for widely spaced poison , r- pins in a wide water gap I13I. This leads to an underprediction of I' poison pin absorption, and an overprediction of K,gg as the [L results will show. j CASMO-3 lattice calculations were performed with both 40 and I m 70 microgroup data. A summary of the Jattice K,(c) is shown in , Table 4.1 The intermediate macro-yeoup calculation in cylindrical L geometry was included in the calcu'.ational Evguence. This s ; C ]'
- 33 -
q o - n [ provides better energy condensation for water gaps and poison
. sheets in the the 2D, 7 group, COXY calculation of the lattice. l I
Four group PDQ data was requested for the fuel, gap and poison r, [ regions. DIXY calculations were perforPed to generate G-factors
,-- for water gaps and fixed poison in order to correct for the V
^ difference between transport theory and diffusion theory i I
p absorption rates. In addition, H-factors were generated for
- L either water or poison. This preserves the transport theory , f lattice reactivity in a diffusion theory calculation I14) i .
CASMO-3 reflector calculation also were performed with both r l U 40 and 70 microgroup data. The reflector calculation proceeds I Ic with micro and macro group calculations, but the COXY calculation h
~-
is simplified to a model of one pin row of the bundle next to the I water reflector. The calculation generates net currents, surface 'b fluxes and region average fluxes in the 2D, 7 group structure. .{e From this data, interface discontinuity factors are determined.
~
These are latter applied to the 4 group water cross sections also rj
" generated in the calculation. Application of the interface discontinuity factors on the water cross sections perserve net ;
leakage at the core reflector boundary I14) Water
' I
. reflector data l
r. was generated as a function of soluble boron only. j {.' e
't L: 4.2 CASMO/ CHART /PDQ Modelling r; CASMO-3 generates a punch file with four group cross section Y
data, G-factors, H-factors and discontinuity factors for each specified PDQ region. This data is processed by the CHART-2 code u l" which applies the various factors and puts the data in the proper l! C
g; .y y w -
^
~
f']
.'. ,K
&. (. * '
- t- ,
.s , .
,9 \ ,
. [' a t .
s v,, U- s a ., N
; ty s y
- c. ,
4 - U> PDQ input format. s \
- ,. N
% T PDQ diffusion theory is used'to mMel the B&W criticalsuiu s
.ge, ,
x - s
,z y,
" Ih.e'X-Y plane. Theeffectofa{ialleak'ageisaccountedfor$y' 'i input axial bucxling of B f 2
4.16 y 2 for Cores I-IX and B,2 = 3.90 , $.c. s < m -2 for Cores X-XXI. This corresponds to cr.4tical water heights e-. - '
. s .
.s adjustud to 145 cm end 150 cm, respectivelf(15;
{ s . The PDQ models for Cores I, IV, VII and VIII were half core x g(. .s ~ N . . , , , L- ' symmetric, and the rest were lower right quarter core symmetric.' i 'i s - ' , s , p ;- A water reflector region roughly '40 cm thici surrounded the cote x s % ,, 1 with tern flux imposed at'the orbr boundary. The steet tan't was '
,. s s ,
[, c. notaddelledsinceitisneutronicallyfa(awayfromthecore. 3 The PDQ models used a mesh spacing of 1; pin pitch in the fuel
- { , , t, and water refledor regions.
m Meshes in the' water gap and poison were selehted to be consistent with the DIXY mer.h +3 sed in
. j'? .
generating G-factors. ForCoresM$,IXandX, this was two y meshes in a water' gap without poisoa shebt The mesh.$'p'achn.g was ,3 x adjusted to account for 6 N increa, sing water gap thickness. s iFor
,7 .N-m.
>Co,:es XI,-XI2I, -a X.TV,'XV, XVII and ...
X?*,, therewastwomeshebinthe pois n"sheet and one mesh on each side of the poisod propar)y off
- .cenceri ng th e s eet. For Cores XII, XVI, XVIII. XX and XXI, there t babtwoNesher-inthepoiron, one mesh on tne narrow gap side and
; s
~
7 . ,, s l two mesher an the vf.de gap side wihh the spacirv,'being at ;a pin "pitch line, .. l, Note, it is important that'1DQ have meshes consistant with or lI '
~
close to the DIXY calculation. The G-factor calcu. lated for the ,
~ , . .s .
j. crons3 sections is very ensitive to thi's assumption. Alan, it N renu.y doUn't matter how fine or coarse e raesh is chosen in bIXY ). ~. ,b .
. s
'h, % . ) , r-x , .
-h's . .,%
\
- = --V . .
~
,F
. D; E because the G and H factor application will preserve the transport theory absorption and reactivity regardless of the mesh spacing.
[Y:' h 4.3 Results and Statistical Analysis Since CASMO/PDQ is a deterministic method, the calculation of h average K,gg and method uncertainty is more straight forward than go from Monte Carlo results. The following standard equations are used: ki g Average K,gg: E = 1*EK g (5) n1 n variance: e = 1 *E(Kg-K) (6) (n-1) 1
? n .
., Bias: AKb= 1*E(Kag-Kg) (7) n1 where:
y Ka g = Measured,K,gg of Critical Experiment i J Kg = CASMO/PDQ K,gg of Critical Experiment i
,, n = Number of Critical Experiments calculated by ,
k-t CASMO/PDQ l i
,. Table 4.2 shows the measured and calculated K,gg(s) for the b~ twenty one criticals. Note, that Cores V-VIII are 1-1.5 toK high I
relative to measurements, these cores, along with IV, contain B 4C
~~
pins in the water gap between assemblies. CASMO la',tice analysis I' was never intended to model this type of situation. However, Core IV is within statistics because the gap is narrow and there are t-
;,; many (84) B 4C in the gap. This situation approaches the BWR n control blade stituation where lattice modelling assumptions would
# l L I
)
..a. , - .
c.- ' ): 't (: R be correct. The statistically analysis to determine the CASMO/PDQ [.7 method bias and 95/95 uncertainty will exclude these cores at the W Spent fuel racks rarely use B C pins for present time. 4 criticality control. However, transport casks sometimes do use {
' such fixed poison and when using CASMO/PDQ in this application, o
V the resultant K bias should be accounted for. {; Thus, excluding Cores IV-VII, the statistics, K i e, for
" CASMO/PDQ with the 70 group library is 1.00280 i .00319 and the bias, AK b, is .00257. .The statistics for the 40 group library
(( are 1.00274 3 00339 and the bias is .00251. For a set of sixteen calculations, the 95/95 one sided tolerance factor is 7 2.524 I9) . This gives a 95/95 method uncertainty for CASMO/PDQ of b .00806 using the 70 group library and .00853 using the 40 group Q. libray b (lG' 4.4 Trends u The core K,gg(s) are organized vs. gap spacing and borated ( aluminum sheet boron loading in Tables 4.3 and 4.4, respectively.
< The results show no trend with either variable.
f .7, c.i. I'b .
- f. .
f ~. Y. l l* 8 -ee 1 C ! l i
-k
\
l
^ ' ^
. . ..= . .
-\
. \
I.
-u I-- Table 4.1 CASNO-3 Lattice K-Infinity mesults ,
C ["' ' Core 40 Group 70 Group i 1.34784 1.34853 p I L II 1.13951 1.13975 )
,. III 1.12194 1.12187 l tw L
IV 1.10316 1.10439 i V 1.09304 1.09402
- VI 1.09304 1.09402 VII 1.09269 1.09320
.P ' VIII 1.09970 1.10025 ' {" IX 1.07969 1.07787 - X ~1.08831 1.08681 J. XI 1.12155 1.12160 XII 1.10529 1.10466 C . l '- XIII 1.10093 1.10165- <L XIV 1.10096 1.10164 l' XV 1.10890 1.10936 p XVI 1.09571 1.09562 XVII 1.11700 1.11733 f.7 L!/ XVIII 1.10094 1.10070
.- XIX 1.12368 1.12384 XX 1.10523 1.10482 XXI 1.09128 1.09018 c:
.' 4' - t.. - IV !bi d L, I, _ i'
1:
- C- .
8 Table 4.2 CASNO-3/PDQ-7 Core K,gg Results c. 1-
. Core Measured 40 Group 70' Group I 1.0002 1.00243 1.00334 j.'j; II 1.0001 1.00325- 1.00363 III 1.0000 1.00622 1.00637 IV 0.9999 1.00382 1.00502 V 1.0000 1.01064 1.01166 VI 1.0097 1.01982 1.02083 Fm N VII 0.9998 1.01422 1.01490 q, VIII 1.0083 1.02265 1.02335 ;; IX 1.0030 1.00192 1.00055 X 1.0001 1.00612 1.00505 XI 1.0000 1.00761 1.00785 XII 1.0000 1.00725 1.00689 F.l. XIII 1.0000 1.00423 1.00502 XIV 1.0001 1.00049 1.00124
. r. , XV 0.9988 0.99556 0.99615
- 1. XVI 1.0001 0.99939 0.99954 XVII 1.0000 0.99918 0.99967 I b
.w XVIII 1.0002 1.00162 1.00143 XIX 1.0002 1.00054 1.00088 [3, XX 1.0003 1.00177 1.00165 , XXI 0.9997 1.00631 1.00560 ', K 1.0002 1.00274 1.00280 n ,
- c; e +.0008
+.00338 +.00319 b '
a m95/95 .00853 .00806 l m AK b .00251 .00257 J--
- Statistics excludes Cores IV-VIII c:
t- , I L f.;. , i p
- - - - , , . . , , - - . . - . . . . . . . - . - - - - - . - . , - - - - - - - - - - _ _ _ - _ _ - - _ _ _ - - - - =
rr I Table'4.3 CASMO-3/PDQ-7 Results vs. Gap Spacing f i spacing (ca) Core 40 Group 70 Group ' 0.0 II 1.00325 1.00363
?:
1.636 III 1.00622 1.00637
!) XI 1.00761 1.00785 l
XIII 1.00423 1.00502 1.00049 1.00124 [' XIV XV 0.99556 0.99615
- XVII 0.99918 0.99967 XIX 1.00054 1.00088
- E 1.00222 1.00272 t i f.
10. 3.273 XII 1.00725 1.00689 f, XVI 0.99939 0.99954 XVIII 1.00162 1.00143
~
XX 1.00173 1.00165 E 1.00251 1.00238
'i i
4.908 X 1.00612 1.00505 r XXI 1.00631 1.00560 f' 1.00533 i
'~
R 1.00622
- l-. ,
k- 6.544 IX 1.00192 1.00055 l i
;~. :
}f ... p E
)
i
,- i t.
]
e, l k .
- i. Table 4.4 CASMO-3/PDQ-7 Results vs. B-Al Sheet Baron Loading c
~
w/o B Core 40 Group 70 Group 0 .1 - XIX 1.00054 1.00088 XX 1.00177 1.00165 XXI 1.00631 1.00560 j; E 1.00287 1.00271 I 0.242 xvii 0.99918 0.99967 1.00162 1.00143 KVIII f E 1.00040 1.00055 F 0.401 xv 0.99556 0.99615 i XVI 0.99939 'O.99954 [; i 0.99748 0.99785 1 1.257 XIV 1.00049 1.00124 l l 1.620 XIII 1.00423 1.00502 l l': (1 c: C. l s ed W N e
r-
! ~
- t. 5.0 APPLICATION OF UNCERTAINTY TO FUEL STORAGE RACK ANALYSIS I
' 5.1 -Criticality Safety Design Criteria The NRC spent fuel storage rack criticality design criteria (8) requires n-
(( E,gg 3 95 with uncertainties. (9) p K,gg is calculated at 95/95 probability / confidence level by the following equation: r- [i K,gg = K,,,+ AKb+ (AKg) + (AK,) (10)
. ;. . where:
K,,, = K,gg for Nominal Configuration, !
,' AK = Calculational Bias, b
AK, = 95/95 Calculational Uncertainty, i
\
_' and AK, = 95/95 Mechanical Uncertainty. , , . , For KENO-Va Monte Carlo, the 95/95 calculational uncertainty l' 1 .L is the root sum of squares of two terms: the 95/95 methodology l
'f uncertainty, an95/95, and the 95/95 KENO-Va uncertainty for the nominal case, i.e.
l AK
-- " IIII c I'm95/95 I + I"95/95*' nom)
T where a95/95 = 95/95 for number One sided Tolerance Factor of generations executed in nominal case, and a nom
= KENO-Va Standard Deviation
- of nominal case.
For CASMO-3 and CASMO-3/PDQ deterministic calculations, the i i second term in Equation 11 disappears and AKc " 'm95/95* j ,. The 95/95 mechanical uncertainty is root sum of squares ) , l
- . - - - + - - a ye-- - , , , , . -a,-, -
,,,-,e- ,,,,,---,-,--,,.--..,,,,,,,_,,_,w,,,,- --
u - a; -
,. a.-. . 'I f r, combination of AK(s) due to mechanical tolerances set at a 95/95 g confidence level, i. e.
AK, = l(e,)2 , g, 3 2 + ('c) + ... (12)
-{..
u where e,, eb' ':, etc are are quantified in sensitivity analysis p of the rack configuration. I.
.The application of the biases and uncertainties calculated for the NITAWL-5/ KENO-Va methodology and for the CASMO-3 methodology is demonstrated below for the Maine Yankee fuel p
L. storage racks. . r ~L; U 5.2 NITAWL-S/ KENO-Va Uncertainty Calculation For the NITAWL-5/ KENO-Va methodology the following biases and uncertainties have been quantified: !
; AK b = .00395 for the 123 group library ; or = .01059 for the 27 group library 4
AK = f r 123 group cales (13) c (.01245)2 + ("95/95*' nom) I.. or
= l(.01155)2 + I"95/95*' nom) for 27 group cales (14) ! .. A KENO-Va 123 group calculation of the Maine Yankee fuel storage racks with 3.25 w/o U235 fuel produced a nominal K,gg and ,_ standard deviation of .88747+.00222. This was the result of 600 neutrons per generation for 200 generations, 120,000 histories.
The 95/95 one sided tolerance factor for the error produced by two [ hundred generations is 1.837. Note, that the one sided tolerance c. factor must be determine by the number of generations executed-l: t, since the KENO-Va standard deviation is based on the number of c, generations and not the total number of neutron histories. The h w K
(.. b, t l' y
" result for this-case is a 95/95 Monte Carlo error of .00408 AK.
? Combining the methodology uncertainty and the Monte Carlo 1.;
uncertainty using-Equation 12, the calculational uncertainty T[ beeones: [ E' AK, = l(.01245)2 + (.00408)2'
= .01310 The sensitivity analysis of the rack unit cell gave a AK,= .00869.
Combining the bias and uncertainties using Equation 10, the 95/95 K,gg becomes: [1 K,gg = .88747 + .00395 + d(.01310)2 + (.00869)2' l'~
= .90714 The value of .90714 is the value with uncertainties to be compared to the NRC limit of .95.
1 E
. 5.3 CASMo-3 Uncertainty calculation
~' For the CASMo-3 and/or CASMo-3/CRART-2/PDQ-7 methodology the
{ following biases and uncertainties have been quantified: i: AK = .00251 for the 40 group library b ; I ({; or - .00257 for the 70 group library
,, and a.- .
= .00853 for the.40 group library l
$~ AK, l j'; or = .00806 for the 70 group library l A 40 group CAsMo-3 calculation of the Maine Yankee fuel storage racks with 3.25 w/o U235 fuel produced a nominal K,gg of l l
.88658. Using the same 95/95 mechanical uncertainty as before, the 95/95 K,gg becomes:
', K,gg = .88568 .00251 + l(.00853)2 + (.00869)2' I l l, 1 l 1 l- - . . , - . . .. . . .. , _ ,.. _ , . . - . . . _ . - _ - _ _ ., ,_____ -_ _ _,_ _ _ _ _ _ l
. 3 I
- r- ;
I.
= .89535
.rr
};, Again, this value is to be compared-to the NRC limit of .95.
4, Al'so, note that this number is lower than the equivalent KENO-Va b K,gg. The differences of the order of 1 to 2 % are to be
~ ~
- g. expected. When discrepancies arise, the higher number is to be used as the limiting value, c:
L'
*e r
1. IL L a 6 im a tg h.' ed V[ l ce m
- 7. .
u e t
- . t 1 1
?
, i:
6.0 CONCLUSION
S f{ k , The YAEC criticality safety methodology was used to model p twenty one B&W fuel storage rack criticals. This investigation was done to validate the methodology for licensing calculations of fuel storage criticality. Analysis was performed with the _ NITAWL-S/ KENO-Va and CASMO-3/PDQ-7. A method bias and method uncertainty for each path was determined from statistical analysis
~
of the calculated K,gg(s). Application of the method bias and
~
uncertainty to the Maine Yankee spent fuel rack' criticality safety analysis was illustrated. The statistics using NITAWL-5/ KENO-Va method on the twenty '% l q one criticals are .99686 1 00525 with the 123 group data and
- q .99051 1 00487 with the 27 group data. The methcd bias and 95/95 ;
j- methodology uncertainty with the 123 group library are .00374 and
.01245, respectively. The method bias and 95/95 methodology
'w. uncertainty with the 27 group library are .01009 and .01155, '] respectively. Trends of K,gg with water gap spacing and boral boron loading are observed and quantified when using the 123 group I .f. library. No trends with gap spacing and borated aluminum boron y, f loading are observed when using the 27 group library, but a trend 238 with enrichment may exist due to the overprediction of U resonance absorption. Th'e statistics using CASMO-3/PDQ-7 method on sixteen of the .: '. criticals are 1.00274 1 00330 with the 40 microgroup data and l
; 1.00280 1 00319 with the 70 microgroup data. The method bias and 95/95 methodology uncertainty with the 40 microgroup data are
[' l u
[' . k r b .00251 and .00853, respectively. The method bias and 95/95 method uncertainty with the 70 microgroup data are .00247 and
.00806, respectively. CAsMo results show less scatter than KENO-Va, and no trends are indicated with gap spacing or borated aluminum boron loading. However, CASMO-3 lattice calculations j r-
[ overpredict K,gg in Cores IV-VIII with widely spaced 54 C pins in a { wide water gap. At the present time, the CA5MO-3 lattice model l ' lq.
- does not have a valid option for these casos. These cores were excluded from the final statistics. In any case, high density
': I spent fuel rack designs rarely have fixed 4B C pins for criticality l
, control. !
a: Application of the biases and uncertainties in the j calculation of K,gg with a 35/95 probability / confidence level was m clearly demonstrated in section 5.0 of report. It is recomended that NITAWL-S/ KENO-Va with the 123 group library and CASMO-3 with the 40 microgroup production library be the standard methods.
~
Each method should be used as a consistency check of the other. l
^; }
Whenever a discrepancy arises, the method yielding the highest i 95/95 K,gg is to be used as the limiting value. However, in the d- application of NITAWL-S/ KENO-Va where wide water gaps are present, f
? underprediction of K,gg should be expected, and in the application of CASMO-3 where widely spaced 4B C pins in water gaps are present, the overprediction of K,fg should be expected.
e
+
-----,-w-,,---w-,,,-- w, --,----
.. d .
~ ,
REFERENCES
- 1. ORNL/NUREG/CSD-2/v2, "NITAWL-S, SCALE System Module for
' h. : Performing Resonance Shielding and Working Library Production",
b ,R. M. Westfall, L. M. Petrie, N. M. Greene and J. L. Lucius, October 1981. P [ 2. ORNL/NUREG/CSD-2/V1/R2, "KENO-Va, An Improved Monte Carlo
. Criticality Program with Supergrouping", L. M. Petrie and
=
N. F. Landers, December 1984.
- 3. STUD 8VIK/NFA-86/7, "CASMO-3, A Fuel Assembly Burnup Program",
User' Manual, M. Edenius, A. Ahlin and B. Forssen, November 'E 1986. g
- 4. YAEC-1453P, "C-H-A-R-T-2 CASMO To RARMONY Tableset Conversion
- Processor User's and Programmer's Manuals", D. Napolitano and P. J. Rashid, September 1984. j
- 5. EPRI/ARMP Documentation ,"PDQ-7/RARMONY User's Manual", B. M.
~. Rothleder, March 31, 1983.
L
- 6. B&W-1484-7, "Critical Experiments Supporting Close Proximity 4
7 Water Storage of Power Reactor Fuel", N. M. Baldwin, G. S. Hoovler, R. L. Eng and F. G. Welfare, July 1979. ; m
- 7. Trans. Am. Nuc. Soc., "Criticality Safety Criteria", W. ,
, Marshall, P. D. Clemson, and G. Walker, Vol 35, pages 278-279, l
- November 1980. ;
l 4' 8. ANSI /ANS-57.2-1983, "Design Requirements for Light Water d.~ Reactor Spent Fuel Storage Facilities at Nuclear Power Plants", i Approved October 7, 1983.
- 9. SCR-607, "Factors for One-Sided Tolerance Limits and for - >
- - Variables and Sampling Plans", Table 2.4, page 46e D. B. Owen, [
March 1963. _. 10. NS&E, "Evaluation of AMPX-KENO Benchmark Calculations for High Density Spent Fuel Storage Racks, S. E. Turner and M. K. 2 Gurly, Vol 80, pages 230-237, 1982. {I"
- 11. Personal Communication, D. R. Harris, RPI, August 4, 1987.
i I 12. EPRI NP-1098, "Analysis of U235_g238 Thermal Reactor l
- Benchmarks Consistency and Interpretation", J. Hardy and D. R.
Finch, Symposium Proceedingst Nuclear Data Problems for [ Thermal Reactor Applications, June 1979.
- 13. Personal Communication, M. Edenius, December 1967.
~
I j~ 14. STUDSVIK/NFA-86/8, "CASMO-3, A Fuel Assembly Burnup Program",
' Methodology, M. Edenius, A. Ahlin and H. Haggblom, November 1986.
L. I l
n.a
- 1 . STUDSVIK/NR-81/61, "CASMO, Benchmarking Against. Critical Experiments in Rack Geometries", M. Edenius, K. Ekberg, E.
Pilat and D. VerPlanck, November 1981. , r.- T:. s.
?
Pg e s.
'I ,
2s Fi
,, l
.a' l i f's I/, a* kl. u l _}}