ML19332C283

Fort St Vrain 3-D Neutron Source Analysis Using DIF3D.

ML19332C283
Person / Time
Site:	Fort Saint Vrain
Issue date:	10/09/1989
From:	Sherman R, Walker V, Warembourg D PUBLIC SERVICE CO. OF COLORADO
To:
Shared Package
ML19332C281	List: ML19332C280 ML19332C283
References
EE-DEC-0022, EE-DEC-0022-R-B, EE-DEC-22, EE-DEC-22-R-B, NUDOCS 8911270096
Download: ML19332C283 (43)

Text

,

q .p y ,,

EE-DEC-0022 ll REV B l

FORT ~ST. VRAIN 3-D NEUTRON SOURCE ANALYSIS USING DIF3D-

\ <

Prepared by: -u hr% 10/n/fy-

'Da t~e Verified by:. /Ith//)&M/in 4/5 9

~

r Date l'

Approved by: 1% Kager '

We -; M-7-b

/ Date i Concur: '

iod /sq  ;

l ProgranfManager Date  !

l Approved by: d N. E , __ _ >- /e-9-#7 D. W. Warembourg p Date 8911270096 DR 891106 ADOCK 05000267 PDC __ ., , _ . _ _ _ _ _ _ _ , , , _ . , _ . , , _ , . , __,,,, _ , _ _ , _ ,

t..

9 TABLE OF CONTENTS

,. i

't Page I. INTRODUCTION 1 II. BACKGROUND' 3

-III.-THE FSV'3-D DIF3D SOURCE MODEL 8~

III.1 Geometric Layout . 8. .;

Input Deck' Modifications

~

1111.2 10 III.3 Calculational Sequence 10' LIV. DIF3D:1/2 CORE MODEL BENCHMARK RESULTS 16:

IV.1) The Experimental Data L1'6 IV.2) Aaalytical Versus Experimental Results 17 V. DIF3D PREDICTED DETECTOR RESPONSE DURING DEFUELING' 21 V.1 Preliminary Design Considerations 21 V.2 Calculational Results 22. -

V.3 Near End-of-Defueling Detector Response for the 232 EFPD FSV Core 25 V.4 Calcu!ational Results Using Only

.the Existing Sources 25

-VI.-CONCLUSIONS / RECOMMENDATIONS !38 VII. REFERENCES 39' Y

L l-li i e

1

w ,

List of Figures i

P,.ag 4 L1). II.1, Schematic of R-Z Model!for Diffusion Theory 6

2) II.2,i Schematic of R-Z Model for Transport Theory 7- ,

3)' III 1, 3-D 1/2 Core DIF3D FSV Model for Detector / Source Analysis 12

4) III.2,-3-D Geometry-in Nodal DIF3D for Fort St.'Vrain 13

.f III.3, Proposed Neutron Source. Locations-to Support 5)

'FSV Defueling 0perations 14 y

L i

. s P

I

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

f.,

.4

, x .. f List of Tables. -

Pagg

-1)-III.1, Methodolcgy for Using the 1/2 Core-Model 15

2) IV.1, Startup Channel

Count Rate Data from the

-FSV Third Refueling- 19

3) IV.2, . Detector Response. Comparisons for 1/2 Core

-DIF3D vs Experimental Data 20 4)'V.1,SourceLocationResults 28-5).V.2,SourceStrength$esults 29 6).V.3, Cocked Rod Results 30

7) V.4, Boronated Dummy Block Replacement Results 31 .

s

8) V.S. Upper Boronated Plenum Results 32

9) V.6, Pure Graphite Source Block Results 33

10) V.7, Near End-of-Defueling Detector Performanee 34

11) V.8, Near End-of-Defueling Performance for the

'232 EFPD FSV Core 35 l-12) V.9, Calculated Flux Using Existing Sources '36 l 13) V.10, Predicted Count Rate at the Start of Refueling 37 l

l, J

[

I-l-

l'

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

6 ~

q y q .

. EE-DEC-0022 L '

l: REV B:

L Page 1 of 39  ;

i

l 3

I.

INTRODUCTION PSC~ has decided to permanently shut down operation of' the FSV 4 plant no later than June 1990.- Defueling of the reactor core is

.now scheduled to begin 100 days after the final core shutdown, f The current plan is to remove the fuel from the core as quickly j as possible by replacing active fuel with boronated graphite.

"dumy" blocks as each core region is defueled.- The use of-boronated dummy blocks raises several- issues with respect to '

core monitoring for inadvertent criticality during defueling.

operations. ,l From a core physics stanopoint, the GAUGE diffusion code is believed to be more reliable in configurations where the active fuel .isleftasmuchaspossibleasonecontiguousportion;.ie,

" islands" of graphite dumy blocks surrounded by- fuei should be dvoided if possible when using GAUGE. To minimize this-problem ,

it was decided that the FSV core would be defueled one region at .

a time by ring from the outer regions- to~ the center. A consequence of this scheme is' that:-the excore f power ' range

' detectors count rate will-probably be lost as soon as the outer l ring of fuel (FSV Ring-4) is removed and replaced with boronated ,

dumy blocks.

A..more important consideration is that the two upper plenum

'startup channel source range neutron detectors (SUC1 and SUC2) will also lose count rates at some point during defueling due to: 1)' loss of core - sub-critical multiplication as fuel is removed 2) loss of the current neutron startup sources in-Regions 15,16, 9, and 22 as these regions are removed 'from the core. Per current plant Technical Specifications a count rate of at least 4.2 counts /second must be maintained in each detector as long as there is a credible possibility of the core '

becoming critical. Plant protection system (PPS) actions in the l startup range may be unreliable if the count rat ( is less than thet value in the base shutdown configuration for SUC1 and SUC2. <

As a practical matter this means that source detector count rate must be maintained throughout most of the defueling period until q criticality is no longer even a remote possibility. i

>d: _ l g

. EE-DEC-0022

' l' REV B -

Page 2 of 39  :

{

It was- decided- that startup - detector count. rate would be

• maintained during defueling by inserting ~one or more edditional' l Cf-252 neutron sources into the core during the.defueling phase.

An analytical model based upon the source option of DIF3D. was developed to support final design of this project. Important:

design parameters include number of. additional sources to use, source- strength, source- location, source block meterial, and- ,

.possible. upper reflector and plenum region modifications. l This paper -documents both.the analytical model-and techniques used as well as the final results and recommendations generated i I as part of the FSV 3-D neutron source analysis design project, j l

1 i

-k r

,- - - - + <

a - - ----n,, . . ,.. ----.,--.--nv~,. v-,. ,,,,a

e .)

.3

' EE-DEC-0022

[ REV B Page 3 of 39 l

II. BACKGROUND ,l

'1 Predicting sub-critical count rates in FSV startup detectors- I SUC1 and SUC2 due to incore intrinsic Cf-'252 neutron sources. is I an extremely difficult task to perform analytically.- A rigorous o treatment of the problem requires a full 3-D mockup of. the core, E upper reflector, upper plenum ' streaming : paths, upper plenum I guide tubes and orifices, and the PCRV concrete surrounding the detectors.. Neutron transport theory is also needed due-to the

~

I many possible streaming paths and the--resultant anisotropic ,

i' '

scattering . behavior- exhibited by neutrons outside the reactor' 1 7 core.. A full. 3-D transport treatment of the. problem was' '

E rejected due to the lack of any such code being available to PSC1 capable of correctly modeling the system.

GA . Technologies has performed all the supporting design' analysis j for sources already located- in the FSV core. Reference 1 1.

documents the original GA analysis used to support placement of i two : sources, one in Region 9 and- the'other in Region 16. for the '

beginning-of-life (BOL) core. A quick synopsis of the calculational methods used at that time, .in 1972-1973, illustrates -how the calculational problem ma f simplified (taken from pp.4-5 of Reference 1)y  : be subdivided and a) A 1-D slab model of the core'was used to_ determine a suitable-axial position for the sources. ,

b) Using the given axial position, use an R-Z diffusion model to determine a suitable radial position-for the . sources.

~ '

Figure II.1 gives a schematic representative. of this:

model, c) For a given radial position, the relationship of discrete sources-(using various numbers and distributions) to annular sources as generated in the R-Z model was studied.

using the GAUGE code "X-Y" model.

d) Once the best location of the sources was established, the detector signal was calculated by:

(1) The leakage neutron distribution out the top of the i core was converted to a rectanDu'lar grid of point sources using the GAMBLE-5 code. The detector signal was calculated by integrating over the source plane using the proper attenuation coefficients for the material between the source plane and the detectors. This was called the Leakage Source Method.

_ _ _ _ _ _ _ _ . _ _ _ . _ _ _ _ _. .~ _ _ _ .._ _ _ _ _ . . _ _ _ _ _ . __ _ _ _ _ . , .

3 ; p LV

~

EE-DEC-0022

'I l: REV-B 1

Page 4 of 39

~

(2) A transport theory R-Z model of the problem was set -

.up, as shown in Figure II-2. Since the model was-much too large to fit on the computer at that time.

-it was subdivided-into a series of layers with each-layer . calculated by 2-D transport' theory and.used as the input boundary condition for:the next.. layer '

in a series of calculations'that finally reached the detector locations. .The TWOTRAN code was used' for this Transport Theory Method,

' e)- Parts .(1) and (2) above have the largest uncertainty due to -the homogenization of complicated- 3-D streaming .c pathways intoL relatively ' simple 2-D models, hence the reason for two independent calculations.

The two main drawbacks of the calculational methods' outlined,  ;

which were certainly understood by GA when the analysis was performed, are as follows:

Use, of the R-Z model results in both annular sources and

~

a)'

annular detectors. Use of GAUGE to judge differences between annular and discrete sources fails to account for l 'the heterogeneous nature of the core in the radial

. direction;- 1e, fuel region RPF's can' vary significantly within a~ ring of fuel. -This'is certainly a- complicating factor in choosing optimal source locations -in the core, b) . Homogenization lof the. upper PCRV above the boronated plenum elements distorts the transport solution at the 1

-detectors. The- Leakage Source Method, although  :

theoretically valid, in practice must be benchmarked to the transport theory results in order to obtain realistic I

attenuation factors. Finally, the TWOTRAN 2-D transport

' solution lat the detectors suffers from the same R-Z model annular region problem, which also must be accounted for.

Use of the 1-D DTF-IV slab geometry model can lead to greater p(Reference 1).roblems if great caution is not exercised by the analyst A:

1

ry .

q j :. - -

^ EE-DEC-0022- -!

.l REV B -

1 Page 5 of 39' I

.l l

l l

An option not available to GA at the time of the aforementioned i analysis was_ to use an explicit 3-D diffusion model when carrying out the source / detector calculations. The DIF3D. code, j

PSC for in-house 3-D fuel accountability- i developed by(Reference 2), also has the ability to perform sub-applications d critical discrete neutron source calculations. PSC decided to l create such a 3-D diffusion model based upon the following i considerations: 1 1

a) An' explicit 3-D model does away with the annular region problem encountered in 2-D R-Z models, thereby obviating i the need to correlate using an additional X-Y model. '

b)- The effects of cocking control rods, removing' various u)per - plenum elements, changing the source location, _

source block material (fuel versus pure t c1anging),the graphite and adding discrete regions of boronated- dummy.

blocks could all be- modeled explicitly at least to the boundary of the~ upper boronated plenum blocks.

l c) A realistic diffusion theory model can reach only to the t upper boronated' plenum boundary due to helium void streaming paths which come after this point.- -This shortcoming in the. model could be accepted for two  ;

reasons. First, benchmark information was available for a variety of configurations ~ during Cycle 4 refueling' in

-which a -new neutron source was deposited in Region 22.

These cases could be used to test the validity of a 3-D diffusion theory source / detector model. Second, the model would reflect the most peaked, heterogeneous distribution of neutrons emerging from the boronated upper plenum blocks. How much the source neutron flux distribution-

" flattens out" as it streams towards the detectors is a matter of some uncertainty. However, it is generally assumed that any flattening of the spatial flux not accounted for in the 3-D diffusion model would result in an increase, not a decrease, in the projected SUC1 and SUC2 count rates.

Based upon the fact that a 3-D DIF3D source model for the FSV core and upper reflector would be unambiguously explicit for the problems of interest and that the projected detector count rates would be conservative, it was decided to proceed with development of the model.

g ,. - _ ..

/' . ,; j r

3 . .- -!

EE-DEC-0022

~ ~

REV B. l Figure 11.1 .Page 6 of 39  ;

Schematic of R-Z Model for Diffusion Theory

.' l-l f

0

- 10.0 REFLECTOR GRAPNITE y, - 20.0

-30.0 .

90AAft0 TOP REFLECTOR 49.4 i

TOP AEFLECTOR 5 0 ,

' 5 Pa

=

. _ _ g- 129.I Z

e g- 166.I CM

, $2 53 54 55

• g- -176.1

$10t REFLECTOR

.' i Zout i Zout 2 20Nt 3 2ONE 4 tout 5 I 367.0 50.3 133.0 219.1 263.9 294.6 403 5 619.6 CM

(

N .' - t EE-DEC-0022  ;

Figure 11.2 REV B .

Page 7of 39 Schematic.of R-Z-Model for -'

Transport Theory

'600.0 I

' 570 7 -

562.5 -

Pcny N*N CONCRETE .Vol0 'OtitCTOR 552.3 -

i 544.0 - 15 ,

539.9 CONCAETE + COOLING Tutt$ 14

'5'37.4 THERMAL SMitLO + LINER 13 534.0 .

.- 1 I

Gulet TV885 + ORIFICES vol0 CM Il 10 .

l t -

i I

341.0  :

. UPeta PLtNUM 12 L -336.8  ; t l soRAfto Tor AEFLECTOR' 8 317.0 y i I

. Top atFLECTot I L 9 l

6 l I 7 i 237.7 200.7 - - g.

• SOURCE 190.0 -

_ 8 g

, 5 5 l- 0 W l

Coat Toe ZONE E S E

I 2 3 4 5 $

E i i l I

g m o e =* en m 4 R O z g d d f~ dd ~~ c, dfi

- ~

j ,T A

l

_ . _ _ _ _ _ _ _ ____. ________._______- _ . . _ . ~ , . . . . _ , . . . . . _ - . - _ - . . . _ . . _ . _ . . . . . . _ _ , _ . . , , _ , , . . , _ y ..- - . . - . . . .

.y. .. . .. _ _ _ _ _ _ _ _ . . _ . _ _ _ _ _ _ _ _ _

, 6.

. EE-DEC-0022 l REV B i

. Page 8 of 39 i

i l

III. FSV 3-D DIF39 Source Model III.1) Geometric Layout

)

The existing 3-D FSV model u.ed for fuel accountability-(Reference 2) was used as the starting point for building  :

the 3-D source / detector model. The existino codel was not - j adequate for source / detector calculatior.s for several l

reasons: j j

a) The full core model did not include upper reflector j boronated plenum elements or include a  !

cciculational space for detector locations. j b) The full core model incorporated ncdal geometry for i which the external source calculational option was J not available. The model would have to be modified to use standard mesh-centered finit.e difference i geonetry.

c) The full core model already uses all the IBM mainfrare CPU space available; larger models incorporating additional upper reflector regions will not fit in the computer. Use of finite difference 3-D models only makes this problem worse.

These problems were xdressed by creating a new 1/2 core 3-D model incorporating reflective boundary conditions at the lower Z axis. The problem space was affectively halved so that additional solution space could be fit in the upper Z direction, including the boronated plenuni and a pseudo-detector r2gion used to count neutrons emerging out of the plenum. The final addition of a finite difference versus nodal solution geometry results in an even larger problem space due to the need for a finer

? axial Z mesh and the increased number of radial mesh solution points. A schematic of the new 1/2 core model is shown in Figure III.1, Note that the diffusion model is very similar to that used by GA for their original R-Z model listed in Figure 11.1. The difference is that the OIF3D model represents the X-Y radial fuel distribution explicitly rather than as annular rings.

3 k

. EE-DEC-0022 f l REV B  !

Page 9 of 39 i l

Figure III.2, a schematic of the current 3-D full core l nodal FSV DIF3D model, is included for comparison to the  !

new 1/2 core model. listed in Figure III.1. Besides the. {

r.odal versus finine difference solution geometry:; i mentioned previously, there are several other differences i worth noting.

l a) Bottom reflector regions 1015 and 1016 are deleted {

from the new model. 1 b) The new 1/2 core model excludes the lower three -l active fuel layers, listed as regions 1 through 507 in the standard model. j c) The boronated side reflector, region 1020 in the- i standard model, is extended over the top reflectot- i region 1018 in the new model, thereby feming the-boronated upper plenum elements. e d) A new pseudo graphite region 1025, composed of the '

same material as reflector regien 1018, was added "

on top of the boronated plenum. Although there in no such real material region in the core, it was added as a means of counting the total neutrons >

exiting the boronated plenum. If one assumes that I the relative flux distribution exiting the plenum 1 L correlates to final detector count rate regardless i cf the streaming paths that follow but are not i included in the model, then taking an edit of total  :

neutrons occurring at the proper radial location of t p region 1025 is a means of predicting SUC1 and SUC2  ;

count rate. .

L . e) The detectors are modeled as separate edit zones of ,

ngion 1025, 30 cm high with a hexagonal cross section having the same area as a fuel bicck. They '

are located to match the actual SUC1 'and SUC2 locations as closely as possible minus the Z direction offset inherent in the model. i f) Potential source block locations are given new l'

DIF3D labels to ensure that only one block comprises the source (remember that the standard DIF3D model has fuel region edits consisting of  :

either One, two, or three separate fuel blocks as  ;

described in Reference 2). Figure 111.3 shows an  ;

l example of existing and hypothetical source block '

j locations. ,

I i

,_+ -.. - . . . . . . . . . - - . - . _ . . . . , - . . . - , _ _ _ _ _ _ . _ _ _ . . _ . _ _ . _ _ - . _ _ _ - _ _ _ _ _ _ . _ _ _ - _ _ _ _ . . _ . _ _ _ _ _ . _ . _ _ _ _ . _ . _ _ . _ . _ _ _

i l

i

~

EE-DEC-0022 j l REV B -

Page 10 of 39

'l l

l 111.2) Input Deck Modifications j The input for the 1/2 core source option is very similar j to that of the standard full core nodal model. The fixed J source option is invoked on card 3 of the A.DIF3D input

-section. The source strength and location are set on card 19 cf A. NIP 3, while special sourca output edit infomation  ;

is set by A. NIP 3 card 40. New boundary conditions, axial J Zmesh,andsolutiontype(finitedifferenceversusnodal) 1 are invoked per the usual inputs in A.DIF3D and A. NIP 3. I i

!!!.3) Calculational Sequence {

- The FSV 1/2 core DIF3D model is able to explicitly model a l variety of core configurations. :The analytical effort of r determing the requirements for the new " shutdown" sources )

involves modeling the following:

a) Changes in source location, radially and axially l:

b) Changes in source strength c) Use of fuel or a pure graphite block to hold the source  ;

d) Cocking or inserting the control rod in the region  ;

holding the source e) Adding boronated dummy blocks.to the core j f) Removing CRD strings out of the upper reflector for fully withdrawn CRD's  : ,

g) Removing selected boronated upper plenum blocks and replacing them with pure graphite.

• The GATZINT/BUGATT codes are used to create fuel and reflector macroscopic cross sections at the burnup time of 1 interest during a cycle. For the purposes of this study, a flat 300 degrees K temperature distribution for the entire system was assumed. Note that a new GATZINT/BUGATT is required for each unique configuration of control rods in tne fuel; ie, for each all-rods-in (ARI) or cocked rod case of interest.

i l

. EE-DEC-0022 .;

I REV B i Page 11 of 39 l i

t The macroscopic cross . section file created by  ;

GATZINT/8UGATT and used as A. ISO section input to DIF3D  ;

can be further modified to accommodate the different cases i listed above (with the exception of cocked rods). . Table  !

III.1 gives an overall view of how the 1/2 core model parameters are set.

1 Typical CPU time runs from two to three and ona-half  !

minutes, which isn't too bad. System resource '

requirements, however, are large. At least 8300K of central processor memory must be used to fit the problem,  !

which is near the limit of the in-house IBM system  !

capacity. The convergence criteria used for this study was 1.E-4,

~

which resulted in convergence after 18 to 35 ,

outer iterations. '

7 l

+

r

-- . , . . . - - , _ . - - - . _ . . - . . . .____.._________-_____._mm _ _ _ - - - - - ~ = - - . - - - - - - ,

_ . . _ ._ __ _ _ _ _ . . ~ . . _ . . _ . . _ _ _ . . . . . _ . . . _ _ _ . . _ . _ . . . . _ _ . _ _ . _ . _ _ _

EE-DEC-0023

• REV B .

Page 12 oT 39 I Figure 111.1 3-D 1/2 Core DIF3D FSV Model for Detector / Source Analysis Cm 367.0 ^

337.0 Graphite " Detector" Material 1025 he'N',"

t 8cronated Reflector 1020 '

317.2 - - -

30 wtt Rodded 40 wt1 Rodded Reflector Reflector (UnroddedReg.1018) (Unrodded Reg. 1018) .

247.9 I 237,g Unrwided Ton Reflector 1017 -

218.1 ,  !

Regions

• Regions i 883-1014 l 846-882 2

! I i 158.6 * *  ! i

. 6

. .b 138.8 E .

. $. k Regions 2 2  !

714-845 l Regions i i ,

8 677;713 3 3  ;

79.3 e r e W  ;

a n e

g -

59.5 $ &  :

Regions I Regions 4 546-676

! 508-544 3 g a 0.0 --------E------- "

i j n A I

l 8 I

i i i

l v v <

Reflective Uncontrolled Controlled 1 Boundary Fuel Regions Fuel Regions Conditions

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

'EE-DEC-0022 i REV B I

- Figure III.2 Page 13 of 39

]

Asial 3D Geometry in Nodal DIF3D for Fort St. Vrain 1.LAs.1 e j 435.8 ., . , , )

30 Wtt Rodded 40 Wtl Rodded i

• ****# * ***** I 16 -

(Unrodded Reg. 9101 1 (Unrodded Reg. 01010) il g  ;;;.; 3. . . ...,..... . >

g lu.0 , ,,

u ions ,i MS -10u Imions i 882 ,

12 456.5 436.7 3

f e

l l Regions l Regions E $ l 11 714 845 t677-713 *

• r 10 371.2 M $

357.4 n

f. C 4 -

I Regions 545-676 Regions 50s-544 f

=

I 297.9 "

g 4 ,

i 278.1 8 ,', In ions 1 Loss $ 1 Y

-507 -375 #

7 l $ l; w ,

218.6 -

i j ,

6 198.8 - -

s -

5 R*f ions I Regions  :

207-330 l 170-306 139.3 "

4 119.5  ;;

3 Regions , Regioat m 38-169 _ __

g 1-37 -

2 60.0 ===--- a==^== == --^=

f 50.0 1 Rotton Reflector 1015 0.0 ,

Region Wanber -

Axial -

30 wtt 40 Wtt a t '

slice Unrodded Rods Rods l i

16 1018 1021 1023 4 1022 1024 "

15 1017 controt, 14 1017 1022 1024 Uncontselled Regions Regsonst

.-e..- +.- s-e-.--w- - - - - - - - . . , - - - ,--.,m-- .-.,e-,----,e--,4

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

EE-9EC 0022 i PROPOSED REV B l NEUTRON SOURCE LOCRTIONS '*****"'

TO SUPPORT FSV OEFUELING OPERRTIONS l Figure III.3 j l

f i

37 20  :

38 i 35 21 8

19 i 18 -

34 22 1 9 '

2 **

7 6 se RJC II  ;

17 23  :

33 10 3a l 1 -

6  ;

32 16 w 24  ;

11  :

4  :

ic L 5  :

suc :  : ** 5 i l 31 25  ;

12 13  ;

14 30 t 26 +

27 28 29 ,

1

(: \

LEGEND .

acoloN tosca 30 704 SOLRCc LOCATIOP6 rom ocnztIts cXISTING SOLRCC LOCATIONS

W EE-DEC-0022 REV B Page 15 of 39 Table !!!.1 Methodology for Using the 1/2 Core Model Modify A. NIP 3 Modify A. ISO Insert Case Macro Deck- Card 14 Card 15 Card 19 Card 30 Label / Macro Upper Ref Dummy Macro Change Source X X X X Location

' Change Source X Strength

' Set CRD X Confi uration in Fu 1 Set CR0 X X X X iConfiguration in Upper Ref i Add /Take Away- X X X X X !

Dummy Blocks >

Change Type of X

Source Block i (Graphite / Fuel) i Remove Boronated X Upper nelenum Blocks- ,

L I

7 w .

r o

1

)

EE-DEC-0022 l REV B l Page 16 of 19 l l

l

)

IV.) DIF3D 1/2 CORE MODEL BENCHMARK RESULTS  ;

IV.1) The Experimental Data I Benchmarking the 1/2 core detector / source model was not a l trivial task due primarily to the lack of clean, '

unambiguous experimental data. FSV site personnel went back to recheck the records for data taken when external sources were placed in the core during previous refuelings  !

at the start of Cycles 3 and 4. Details of this task a re .

documented in Reference 3. Count rate dsta taken from the  !

beginning of Cycle 3 was judged to be unreliable dee to a l number of apparent incor.sistencies in the records. Data  :

from the beginning of Cycle 4 wh2n a new source was placed  !

in Region 22, was judged to be acceptable and fonned the .

basis for the subsequent benchmark calculations, t Table IV.1 lists the Cycle 4 data, taken from surveillanc.e  !

records and strip chart recordings, for startup detectors ,

SUC1 and SUC2 during the third refueling. Although the  :

surveilip ~ e data is believed to be the most accurate, the  ;

strip ch d L cases. form the only basis on which to test the models for CRD's and boronated plenum blocks in tha upper  ;

reflector.

Inspection of this experimental data reveals several  ;

interesting results: '

a) The insertion of the source in Region 22 had no- l significant effect on the detector located on the otner side of the core. The SUC1 count rate went ,

from 31.3 to 32.2, which is within statistical experimental error, b) Between the beginning and end of refueling, sfter the source went into region 22, the count rate of both detectors increased slightly, from 32.2 to ,

35.0 and from 52.6 to 56.0 respectively. This -

increase could perhaps be attributed to the overall increase in core reactivity due to the insertion of i l four fresh fuel regions. .

-. ,._n. - , . .....-.._,,g ,n.., ,,n-.n , . - . . , .

rs I

i 4

e-EE-DEC-0022 l REY B Page 17 of 39 c) The strip chart result of 90 counts /sec for SUC2 after the insertion of the source in Region 22 is in significant disagreement with the surveillance 1 data showing 32.6 counts /sec for the same case. J Conflicting experimental results such as this show  !

that caution must be exercised when comparine analytical model results to experimentai j

" benchmarks". -

m  :

e i

IV.2) Analytical Versus Experimental Results i Table IV.2 lists a comperison of DIF3D total neutron flux  :

edits versus experimental startup ' detector count rate.  :

l Notti that the compart:;ons for the strip chart data are net ,

exact because the number densities used by DIF3D were for i all regions refueled, while in reality only regions 3 and  ;

22 had been refueled when this part of the data was taken.  ;

i The analytical results compare quite well to the measured  ?

l data. The relevant test is the unifomity of' the f '!

neutrons / detector count parameter. If this parameter was ,

a constant for all cases, then the analytical model could ,

be applied with a very high degree of confid'r ce. >

The surveillance cata was judged to be the most accerate  !

(Reference 3). For this case there is near unifomity between detectors when comparing -measurement and prediction, with about 20 neutrons needed per detector count.. Unfortunately, the strip chart data was available -

for only one detector (the one closest to Region 22 where .

the new source was inserted). For hese cases the range of f neutrons / detector count went from 21.2, which agreed with the surveillance results, down to 7.98.

L Based epon these benchmark results several important i observations can be made:

a) Comparison of predicted to measured count rates is judged to be good. There is no more than a factor of three difference in measured versus predicted counts over all the benchmarks analyzed. ,

--c.. , . , . , , . - - - . . . - , . . - , , - - - - , - - - . . . - - - - , - - - - - - -- - - - - - - - - - - - - - - - . - - - - - - -

.* l

< l i

EE-DEC-0022 .

l REV B l Page 18 of 39 l t

I b) Although the total number of benchmark cases was small, at least one (the case with no upper plenum blocks and the i CRD removed from the system) tested the model for a fairly  :

unusual core configuration. The model was apparently  !

adequate for this case, giving added assurance that  ;

unusual core configuratins that arise - during defueling i (such as the addition of boronated dummy blocks .re'noval 1 of plenum elements, CRD's in the upper reflector, e t c'. )

will be analyzed reliably.

1 c) from the benchmarks it appears that using a 22 )

neutrons / count conversion of the DIF3D solution flux for '

predicting detector counts during defueling fonns the best 1 basis for ditsigning new external sourec:. The data shows '

no instances in which the i neutrons / detector count ratio.  ;

is any larger than this. The data does show, however, that this ratio could in some instances be significantly '

smaller than 22. If this happens it will increase the actual countrate over what wss predic h:d using 22 +

neutrons / detector count, and thereby increase detector efficiency in a beneficial manner. The design concern is maisetaining at least a minimum count rate during defueling. Getting more counts that expected through a decrease in the # neutrons / detector count ratio would be a

bonus. ,

l d) Using an empirical 22 neutron / count ratio .'or predicting i detector counts would also be conservative if the ,

p , dispersion effect of streaming neutrons outside the plenum i D l is significant; i.e., thr radial flux distribution at the -

I actual detector locatwas is less peaked than it is as it l l- exits the boronated plenum. The analytical predictions

' using DIF3D should therefore contain an additional level ,

of conservatism by not taking any credit for flattening of i the radial neutron flux above the upper boronated plenum, i

e m-

' * ~ a-W -- m w- e we-*w*gui's e-e-w "w++ ret e- a w eawe- w'et's 9 --r=-rw'y m-+-e--t+ --p*- e e t-t g yy- g

.. l EE-DEC-0022 i k l REV B  ;

Page 19 of 39 1 I

Table IV.1 Startup Channel Count Rate Data From j the FSV Third Refueling

1) Surveillance Data 1

Core Configuration SUC1 (CPS) SUC2 (CPS) l Initial (Prior to Refueling) 31.5 7.3 ,

l Prior to Refueling Reg 3 32.0 7.9 i Prior to Refueling Reg 22 31.3 7.9 i Prior to Refuelin Reg 18 32.2 52.6 Prior to Refuelin Reg 29 34.3 56.1 -

Prior to Refuelin Reg 3 34.6 56.8  ;

Prior to Refueling Reg 33 37.5 57.0 Final (AfterRefueling) 35.0 56.0 j

2) Strip Chart Data 1 Core Confifluration SUC1 (CPS) SUC2 (CPS)

Af teFTJurce p' aced in Reg 22. Two Layers of Not Available 1600 .

Reflector Blocks added but '

no Plenum Blocks or CRC in Reg 22

!- After Source placed in i Reg 22. Top Reflector and Not Available 320 Plenum put back, no CRD in .

o Reg 22 '

l Af ter Source placed in

Reg 22 Top Reflector, Not Avaitable 90

. Plenum and CRD put back in L Reg 22 L

[

V l

i p *:

2E-DEC4022 i

! REV B Page 20 of 09 l

Table IV.2 j 1 Detector Response Comparisons for 1/2 Core DIF3D vs Experimental Data j

1) . Surveillance Data: All Regions Refueled After Source in Region 22 {

1. Neutrons / . # Neutrons /

SUC1 SUC1 Count SUC2 ~ SUC2 Count  ;

Measured CPS 35 56 l DIF3D Total Neutrons 6.68E+2 19.1 1.28E+3 22.9

2) Strip Chart Data: Source in Region 22, No Plenum in Upper Reflector l CRD Removed from Core  :

1. Neutrons / # Neutrons /

SUC1 SUCI Count SUC2 3UC2 Count Measured CPS Unavailable 1600  ;

DIF3D Total Neutrons 6.82E+2 Unavailable 3.39E+4 21.2 s

3) Strip Chart Data: Source in Region 22 Plenum Blocks Put Back in Core, CRD Removed from Core .

1. Neutrons / # Neutrons / l SUCI SUCI Count _,SUC2 SUC2 Count Measured CPS Unavailable 320 DIF3D Total Neutrons 6.82E+2 Unavailable 2.50E+3 7.8 t

4) Strip Chart Data: Source in Region 22 Plenum and CRD Back in Core  ;

1. Neutrons / # keutrons/

SUC1 SUC1 Count SUC2 SUC2 Count Measured Ces Unavailable 90 DIF3D Total Neutrons 6.68E+2 Unavailable 1.2BE+3 14.2 ,

ws ...,,.w., . , . . ,,, , . - - . - . - . . - . ~ . . . , . . . . . .

7,

,e 1

EE-DEC-0022  !

l REV B I Page 21 of 39 l

)

-)

l V.) DIF3D PREDICTED DETECTOR RESPONSE DURING DEFUELING V.1) Preliminary Design Considerations

{

Aseries of--DIF3D source calculations were performed in i order to assess in-core source requirements during l defueling. The primary constraint was to have at least 10 l' counts /second on each detector until the core has been defueled to a point at which there is no credible chance of the system becoming critical. At this time it is '

projected that a count rate must be maintained until only the last 9 out of 37 fuel regions remain. At this time '

the all-rods-out (ARO) KEFF of the core will be less than O.95. This design constraint mrf -hange in the future.

However, it is obvious that a nt rate must be lost at i some point during the c'efueling and the ARO 0.95 KEFF  ;

limit seems to be a reasonable point at which maintaining

• counts no longer is justified.

As a consequence, any in-core sources inserted for [

defueling must be located in regions that are to be  !

defueled very late in the process. For this reason it was  :

decided to concentrate on Ring 1 (Region 1) or Ring 2 (Regions 2-7) locations to hold the new sources, ,

r Preliminary calculations showed that putting just one new  !

source in the center of the core would not deliver the necessary sta rtup channel count rates unless the source  ;

were to be made unrealistically large. Count rate was  :

seen to definitely increase as the source was movet.

outward from the center of-the core towards the same X-Y plane location as the detectors. As a starting point, therefore, it was decided to model two separate sources. -

each located in Ring 2 as close to the detectors as possible (see Figure III.3).

The effect of burnup on count rates was handled by  ?

modeling each case at two different burnups, 250 EFPD and 475 EF00. 475 EFPD represents an absolute maximum burnup for FSV Cycle 4 plus coastdown, and yields the poorest detector response for a given source. 250 EFPD results allow for assessment of change in expected count rate with  ;

burnup and would be a conservative predictor of count rate '

if the plant were to be shutdown befon 250 E7PD was achieved.

5 1

y - _ . . - , _ - , - - - , , , , . , _ . _ , _ - , . , , , , , , _ . , _ ._~.__a___.____.

( wa .

!'.. l i

EE-DEC-0022 i l REV B  :

J Page 22 of 39 j i

The sensitivity analysis perform 2d involved a series of two source calculations performed at 250 and 475 EFPD of  :

Cycle 4. As an extra conservatism, all existing in-core  !

sources (Figure III.3) were removed from the mudel. At  !

the point when counts drop the lowest, after Ring 4 and. l most of Ring 3.has been defueled. these sources would not t be present anyway. The dummy block macroscopic cross .

sections are the same as those developed by GA for use in i the 2-D GAUGE code model (Reference 4).

V.2) Calculational Results The DIFJD model source / detector results for a variety of  ;

defueling core configurations are listed in Tables V.1 through V.6. From this data a number of source / detector .

design parameters can be assessed:

a) Source Location - Count rate is maximized when the detectors are located as close radially to the .;

detectors as possible'as part of the top layer of  :

fuel. Tabit.s V.1 and V6 clearly indicate that  !

Region 3, Column 3 and Region 6. Column 6 (Figure III.3) yield the highest counts, b) Source Strength - Count rate increases linearly

'wTth increases in source strength, as shown in Table V.2. From an overall assessment of the results, a strength of 4.0 E+9 counts /sec of CF-252 i for each source was Judged to be the best choice to both yield acceptaMe detector performance and be ,

within acceptable limits for hardling by plant personnel.  :

c) Cocked Rods -

The count rate increases due to cocking the control rods in the regions containing ,

the new sources was found to be quite marginal. '

Table V.3 shows that a maximum increase of only 26% ,

in either detector occurs when the rods in Regions 3 and 6 are withdrawn at the same time. The decreased margin for rod withdrawal accidents '

during refueling due to these rods being out of the core probably makes this an unviable option for '

increasing detector counts.

i

.. i

, l

}

,o ,

i --

EE-DEC-0022 i s

4 l REV B l Page 22 of 39 l t

, j U- .t r

d) Re lacement Boranated Dumy Blocks - The results s ed in Table L 4 show tha". the loss of detector  ;

counts due to insertion of boronated dummy blocks l into the core is very significant. If no other i means are used to increase count rate, then j response becomes clearly unacceptable after Rings 3 ,

and 4 are defueled. Only marginal acceptability  !

exists for the Ring 4 only defuelad cases whkh i fall right on the 10 counts /sec administrative t limit. As noted previously, however, it is not f anticipated that startup channel count rate will need to be maintained up until all of Rings 3 and 4  :

are defueled.

e) Replacement of Upper Plenum Boronated Elements - A ii&nparison of the results listed in Tables V.4 ind  !

V.5 indicate - that the removal of six of the se.ven  :

boronated plenwn elements (leaving the central CRD element) in Regions 3, 6, 10, and 16 results in a i significant boost in count rate. Counts would be (

expected to increase at least by a fator of two in ,

each detector, mostly due to the replacements in Regions 10 and 16. l f) Pure Graphite. Source blocks - Use of a graphite  ;

FeTlector t, lock to hold the proposed Cf-252 surces results in marginally increased de9ctor response, as shown in Table V.6. Since there is no penalty .

for using a graphite b1cck to hold the source. it is recomended that this be done rather than use an i unirradiated Segment 10 fuel block for task.

Based upon the insight gained from the above analytical results, one last series of calculations were performed. ,

The purpose was to show that sufficient counts are maintained as long as required during defueling, even for  ;

i 475 EFPD core, as long as the core is configured properly. The results are listed in Table V.7. Note that only nine rcgions remain to be defueled, at which point analysis shows that the core KEFF will be-less than 0.95 ,

with all rods out. This data shows that 4.0E+9 sources in i Regions 3 and 6, in conjunction with replacement of  :

boronated plenum elements in Regions 10 and 16. meets startup detector count rate requirements. This a nessment takes no credit for other factors which should resu~ t in even higher count rates:

I'

N; l l

.. 3

~

EE-DEC-0022 l REV 8 i Page 24 of 39

]

i i

a) A final FSV burnup of less than 475 EFPD. l o) Replacement of boronated plenum with a non. l baronated version of the same material in Regions 3 '

and 6 along with Regions 10 and 16, i c) Procurement of startup detectors having greater sensitivity tnan the existing ones.

d) Increasing the " gain" of the detectors to levels greater than what was used in obtaining the beginning of Cycle 4 benchmark data.  ;

e) The buildup of intrinsic core neutron sources from  :

the beginning of Cycle 4 until shutdown, due mainly to Cm-242 and Cm-244, would tend to increase counts .

during defueling but were neglected in this study,  !

r f) The analytical m< del, which neglects dispersion of l l the radial neutron flux streaming out of the upper plenum, is believed to be conservative. .

g) Taking no credit for the existing neutron sources before they are removed from.the core.

One non-conservatism- in the preceding analysis was the l l: neglect of the 2.6 year half-life of the Cf-252 sources. '

! In the past, during cycles 1 and 2, the startup detector I

count rates dropped faster than the Cf-252 half-life due >

l'l to deterioration of the startup detectors themselves. GA ,

claims that the deterioration half-life was about 18 months. The standard PSC procedure, however. is to e replace defective detectors when performance noticably ,

degrades. In fact, since the third refueling (Feb.1984) detector count rate hus shown no apparent deter f oration. .

Channel 1 and Channel 2 count rates have been maintained  :

up to the present time at about 30 counts /second by increasing voltage and discriminator levels on each detector and/or by replacing the detectors themselves,  ;

i thereby fully compensating for the decay of C#-252 for a period of five years. Since the ' anchmark calculations '

were done at the third re/ueling when detector voltage / discriminator levels were set at levels lower than at the present time, it is believed that the same margin t l exists during defueling for at least five years to fully compensate for Cf-252 decay. If this is not the ecse, then a new set of replacement detectors should be useo which have at least the same sensitivity at, exhibited by the current startup detectors frca 1984 to 1989. For this

o

}

EE-DEC-0022 l REV B -

Page 25 of 39.

I I

reason and the factors listed above, it is believed that the predicted count rates presented are conservative.

l V.3 Near End-of-Defueling Detector Response for the 232 EFPD l FSV Core  ;

After it was . apparent that Cycle 4 at FSV would end at '

232.0 EFPD, additional cases were run at 250' EFPD using the actual core configuration anticipated near the end of I defueling. 250 EFPD was .hosen because it allows for direct- comparison to the other results listed and is conservative with respect to count rate when compared to  !

the actual 232 EFPD core. The results are listed in Table V.8. These data indicate the following:

a) Adequato count rate is maintained at all times ,

during defueling for the 232 EFPD core.  !

b) When all rods are withdrawn with only nine fual .

regice,. remaining, the operators will be aware of l

an expected SUC1 and SUC2 detector response.

l V.4 Calculational Results Uring Only the Existing Sources A second series of calculations was perfomed after the preceding analyses using only the current existing neutron sources as the basis for the source model. The reasons for perfoming these additional calculations were twofold:

o a) The results would indicate how many core ."uel regions could be removed before detector count rate

. drops below 10 counts /second if no new sources are inserted. This would allow for greater flexibility in scheduling new source delivery if the results t show that defueling could proceed for an ,

appreciable length of time without the new neutron -

sources. ,

3 p

O i O

[ -

.! 1

., i

. EE-DEC-0022 i l REV B j y Page 26 of 39 i t l

)- .!

.i b) The degree of conservatism in the calculations due l l to using beginning-of-cycle 4 detector efficiencies could. be assessed by decaying the present in-core i sources to July 1, 1990 and redoing the previous  ;

analysis using the same 22 neutron / count ratio. i Since at the present time, June 1089, a count rate ]

of 6 bout 30 counts /second still exists for each  ;

detector, calculational results predicting l significantly lower count rates would illustrate i the importance of discriminator /voitage settings on measured count rate and provide an estimate of how conservative the preceding calculations are in not  ;

taking ar.y credit for detector efficiencies greater ,

than those in effect at the beginning of cycle 4. '

The calculational results using the existiny neutron sources as '

l of July 1, 1990 are presented in Tablas V.9 and V.lu. Table V.9 lists the total neutron flux present at each catector as calculated by DIF3D. These numbers are relatively low- in -

comparison to those calculated for the new source phase of this  :

l study. Table V.10 correlates this flux to predicted detector '

count rates assuming: 1) that tha detectors are set to the same beginning-of-cycle 4 efficiency as was used for the new source analysis, 22 neutrons / count and 2) a 30 count /second count rate -

l can be maintained for each startup channel until the start of .

defueling(7/1/89)byincreasingdiscriminator/voltagelevels. 'l l The assumption that the 30 counts /second response of each startup detector can be maintained for at least another year 1- appears to be quite likely given the past performance outlined  ;

in the previous section for the time period from r-bruary 1984 to the present, in addition, site personnel have estimated that only 60% of the available discriminator / voltage gain have been  ;

used ever this period, leaving an additional 40% margin for the i next year of operation plus defueling. Given the validity of this assumption, the results presented in Table V.9 are  ;

significant for two reasons:

l a) Detector count rate is shown to be sufficient u using the existing sources for the first 15 L fuel regions to be renioved from Ring 4 ,

l Therefore, starting e fueling without the l proposed new sources in Regions 3 and 6 ,

I appears to be a very viable option.

Calculationally, count rate is not lost until ,

Regions 22 and 31 are removed from the core ,

using the proposed defueling sequence.

, .,-,nr.

3 5

)

EE-DEC-0022  :

.l REV B  :

1 Page 27 of 39 b) The -margin of conservatism ascribed to using the discriminator / voltage settings in place at i the beginning of cycle 4, rather than what- .;

will be in effect at the start of defueling,  :

can now be better quantified. It appears ~ that  !

this effect will be at least ~ as large as a factor of three (30/9) at the start of '

defueling (i.e., if the detector settings are j not reset, then~ the predicted count rates given for new sources inserted on 7/1/90 ,will t

, underpredict counts by at least a factor three  :

due only to differences in detector " gain"). ^;

This margin should be able to compensate for -

radicactive decay of the Cf-252 for. the entire. ,

defueling period. Furthermore, any additional margin that exists at the start of defueling '

(the present 40% minus that used between now and 7/1/90) will also .be. available .to '

compensate for source. decay and detector degradation.  :

I e

t

o.

r l' EE-DEC-0022 )

l REV B r Page 28 of 39 l Table V.1 Source Location Results i Assumptions: (1) Source Strength = 4.0E+9, (2) All Rods In (3) No Boronated Dumy Blocks Inserted

]

1)' 250 EFPD SVC 1 SUC 2 Source Predicted Predicted Case Locations # Neutrons Count # Neutrons Count 1 Reg 3 Col 3 7.10E+2 32 8.44E+2 38 Reg 6. Col 6 ,

2 Reg 3, Col 3 3.25E+2 15 8.54E+2 39 Reg 5 Col 7 3 Reg 3, Col 3 3.50E+2 16 8.51E+2 39 Reg 5, Col 6

2) 475 EFPD SVC 1 SUC'2 5 Source Predicted Predicted '

Case Locations #Neutrais Count # Neutrons Count 1~ Reg 3 Col 3 6.16E+2 28 7.34E+2 33 Reg 6, Col 5 2 Reg 3, Col i 2.67E+2 12 7.41E+2 34 Reg 5, Col /

3 Reg 3. Col 3 2.88E+2 13 7.38E+2 34 Reg 5, Col 6 r ,., - - ,  %~<. - , - - , , - -

[

• i EE-DEC-0022  :

E l REV B Page 29 of 39  ;

1 Table V.2 {

Source Strength Results ,

Assumptions: (1) Sources Located in Reg 3, Col 3' and Reg 6. Col 6 (2) All Rods In (3) No Boronated Dommy Blocks

1) 250 EFPD SUC 1 SUC 2-Source Predicted Predicted -

Case Strengths # Neutrons Count kNeutrons Count .

1 2.0E+9 3.54E+2 16 4.21E+2 19 2.0E+9 -

2 4.0E+9 7.10E+2 32 8.44E+2 38  !

4.0E+9  :

3 6.0E+9 1.06E+3 48 1.27E+3 58  !

6.0E+9 ,

2) 475 EFPD SUC 1 SUC 2 Source Predicted Predicted Case Strengths # Neutrons Count # Neutrons Count 1 2.0E+9 3.07E+2 14 3.66E+2- 17 2.0E+9 l 2 4.0E+9 6.16E+2 28 7.34E+2 33 L 4.0E+9 ,

1 3 6.0E+9 9.23E+2 42 1.10E+3 50 6.0E+9 4

I I

l

EE-DEC-0022 l REV B Page 30 of 39 L

Table V.3 i Cocked Rod Results Assumptions:~(1) 4.0E+9 Sources in Reg 3, Col 3 and Reg 6. Col 6 (2) No Boronated Dummy blocks Inserted l (3) All Rods in Except Those Specified

1) 250 EFPD SUC 1 SUC 2 CRD Withdrawn Predicted Predicted Case in Regions _ . # Neutrons Count #Nettror.s Count 1 3 7.20E+2 33 1.07E+3 49 2 6 8.32E+2 38 8.50E+2 39 3 3+6 8.46E+2 38 1.08E+3 49

2) 475 EFPD SUC 1 SUC 2 CRDS Withdrawn Predicted Predicted Case in Regions # Neutrons Count # Neutrons Count 1 3 6.22E+2 28 9.02E+2 41 2 6 7.05E+2 32 7.37E+2 34 3 3+6 7.12E+2 32 9. 08E+2 41 e

l

c; . l l  !

{ *' j EE-DEC-0022 i l REV B Page 31 of 39 I t

Table V.4  ;

Boroaated Dumy Block Replacement Results Assumptionst. (1) 4.0E+9 Sources, Located in Reg 3. Col 3 and Reg 6, Col 6  !

l (2) All Rods In for Fueled Regions

-(3) All Rods. Out for Defueled Regions  ;

1) 250 EFPD SUC 1 SUC 2 i

1. Regions Predicted Predicted i Case Defueled # Neutrons Count # Neutrons Count  :

1 1 18(Ring 4) 2.22E+2 10 2.44E+2 11  ;

2 30(Ring 3+4) 4.08E+1 2 4.46E+1 2  ;

3 26(Ring 3+4) 1.98E+2 9 2.16E+2 10

- (Regs9,10,  !

15,16)

2) 475 EFPD SUC 1 SUC 2

1. Regions Predicted Predicted .

Case Defueled # Neutrons Count # Neutrons Count .;

1 18(Ring 4) 2.05E+2 9 2.24E+2 10 1 2 30(Ring 3+4) 4.02E+1 2 4.41E+1 2 3 26(Ring 3+4) 1.86E+2 8 2.01E+2 9

- (Regs 9,10 ,-

15,16)

n.: .

)

i

. i EE-DEC-0022 i l REV B i Page 32 of 39 )

Table V.5 Upper Boronated Plenum Results  !

Assumptions: (1) 4.0E+9 Sources, located in Reg 3, Col 3 and Reg 6 Col 6 (2) All Rods In for Fueled Regions (3) All Rods Out for Defueled Regions (4) 6 of 7 Plenum Elements Replaced in Specified Regions (Central CRD Column Glock Not Replaced)

(5) Rings 3 + 4 Defueled

1) 250 EFPD SUC 1 SUC 2

• Plenums Replaced Predicted Predicted 6

. Case in Regions # Neutrons Count # Neutrons Count ,

1 10 + 16 9.35E+1 4 1.02E+2 5 l

2) '475 EFPD SUC 1 SUC 2-Plenums Replaced Predicted Predicted Case in Regions # Neutrons Count # Neutrons Count 1 3+6 4.08E+1 2 4.48E+1 2 I

2 10 + 16 9.21E+1 4 1.01E+2 5 3 3 + 6 + 10 + 16 9.64E+1 4 1.06E+2 5 f

a 1

--r . , - . , . . . . . , , _ _ e _ ,..-,____,_..m_, . . ,..m. . .._. ,

~

EE-DEC-0022 l REV B-Page 33 of 39 i

Table V.6 l Pure Graphite Source Block Results Assumptions: (1) Use a pure graphite block to hold 4.0E+9 sources, rather than a standard fuel block.  !

(2) Model various core locations: .

1 = Reg 3, Col 3; Reg 6. Col 6; Top Fuel Layer 2 = Reg 3 Col 1; Reg 6, Col 1; Top Fuel layer

.3 = Reg 3 Col 3; Reg 6. Col 6; 2nd Fuel Layer i frum Top i (3) All cases at 475 EFPD )

SUC 1 SUC 2 Core Predicted Predicted Configuration Location # Neutrons Count # Neutrons Count All Rods In  !

1 6.50E+2 30 7.48E+2 34 No Dummy Blocks 2 2.'25E+2 10 2.91E+2 13 .

3- 2.56E+2 12 3.22E+2 15 Regions 3+6 1 7.59E+2. 35 9.33E+2 42 Cocked i No Dummy Blocks 2 4.12E+2 19 6.12E+2 28-3 3.52E+2 16- 5.03E+2 23  ;

All Rods In. 1 2.18E+2- 10 2.30E+2 10 .

Ring 4 Defueled 2 6.69E+1 3 8.05E+1 4 -

3 4.91E+1 2 5.66E+1 3 Regions 3+6. 1 2.48E+2 11 2.76E+2 13 '

Cocked Ring 4 Defueled 2 1.19E+2 5 1,62E+2 7  :

3 7.08E+1 3 9.49E+1 4 '

All Rods In, 1 3.6GEo1 2 3.84E+1 2 Rings 3+4 Defueled 2 8.39E+0 0 1.07E+1 0 3 1.81E4 0 2.52E+0 0 Regions 3+6 1 3.80E+1 2 4.09E+1 2 Cocked.

Rings 3+4 Defueled 2 1.31E+1 1 1.84E+1 1 3 2.79E+0 0 4.41E+0 0 l l

l i

J._..... . _ , . ._ _ . . . . . . .. . . _ , _ _ , _ . _ _ . . - . _ .

~

< l 1

. EE-DEC-0022

.l REV B  !

Page 34 of 39 .i 1

j

~

Table V.7  ;

Near End-of-Defueling Detector Perfonnance  !

i Assumptions: (1 4.0E+9 Sources in Reg 3, Col 3 and Reg 6, Col 6  :

(2 Rings 3+4 Defueled Minus Regions 10 and 16 3 All Rods In for Fueled Regions .

l 4 All Rods Out for Defueled Regions '

5)475EFPD +

1) Regular Fuel Block Sources SVC 1 SVC 2 Cere Predicted Predicted Configuration # Neutrons Count # Neutrons Count Regular Reg 10+16 Plenum 1.17E+2 5 1.38E+2 6 Replace Reg 10+16 Plenum 2.58E+2 12 3.01E+2 14

2) Pure Graphite Source Blocks SVC 1 SVC 2 -

Core Predicted Predicted Configuration # Neutrons Count # Neutrons Count Regular Reg 10+16 Plenum 1.17E+2 5 1.35E+2 6 Replace Reg 10+16 Plenum 2.60E+2 12 2.94E+2 13' i

l 4

i

. -.. , . , -.. -- - ,-.,a-e, . - , , - - .

?

'o EE-DEC-0022 l REV B

.Page 35 of 39 l Table V.8

. Near End-of-Defueling Performance for the 232 EFPD FSV Core-I- Assumptiens

4,0E+9 Sources in Reg 3. Col 3 and Reg 6. Col 6 l Rings 3+4 Defueled Minus Regions 10 and 16  !

250 EFPD Burnup Use of Pure Graphite Source Blocks ,

(S Replace Boronated Plenum Elements in >

Regions 3, 10, 6, and 16 SUC 1 SUC 2  !

Core Predicted Predicted i Configuration # Neutrons Count # Neutrons Count l All Rods In For Fueled 2.78E+2 13 3.17E+2 14 Regions All Rods Out for 1.52E+3 69 3.16E+3 144 All Regions l

D L

.h l '

I s

...n.. - - - , ,. ..n, , a-,- .,. . - , -_-,-y..,.,

-g.

1 . ,s* a; i

c-4 -I [J )

~

EE-DEC-0022 >

l- REV B .

+

Page 36 of 39  :

4  ;

.i Table V.9 d'l Calculated Flux Using Existing Sources Assumptions: (1) Use Existing Sources;in Regions 9, 2215,  !

and 16 Deceyed to 7/1/90  !

(2) All Rods In  ;

250 EFPD 475 EFPb

1. Neutrons # Neutrons # Neutrons # Neutrons -

Coro Configuration SUC1 SUC2 SUCI SUC2

No Defueled Fegions, 90.5 191.4 82.7 179.9 '

4 Old Sources- . ,

Two Defueled Regions 71.2 156.5 66.1 148.6 "

(Reg 23and32),

4 Old. Sources- ,

>r 15.Defueled Regions 68.9 138.9- 64.6 133'4 . 4 (Reg 22and-31 Remain),

p 4_Old Sources o

Ring 4 Defueled, 25.3 L1.8 24.3 1.7

-3 Old Sources 1-

!'L i.

/v

[ .

r b

i

g

- ' f s

.v:

r

, EE-DEC-0022 l REV 8

, Page 37'of 39 Y

l Table V.10 Predicted Count Rate at the Start of Refueling:

Assumptions:-(1) Use Existing Sources in Regions- 9, 2215, and 16 Decayed to 7/1/90 (2)AllRodsIn

1) Use BOC 4 Detector-Efficiencies. 22 Neutrons / Count 250 EFPD 475 EFPD Predicted Count Rate Predicted Count Rate Core Configuration- SUC1 SUC2- SUC1 SUC2 No Defueled Regions, 4 9 4 8 4 Old Sources Two Defueled Regions. -3 7 3 7 4 Old Sources 15 Defueled Regions 3 6 3 6

, 4 Old Sources Ring 4 Defueled, 1 0 1 0' 3 Old Sources

+

2) Normalize Count Rate to 30 Counts /Second on-7/1/90 250 EFPD 475 EFPD Predicted Count Rate Predicted Count Rate foreConfigurathn SUg SUC2 SUC1 SUC2 1

? Ra Defueled Regions, 30 30 30 30 4 Old Sources Two Defueled Reglens, 24 25 24 25  !

4-Old Sources 15 Defueled Regions 23 22 23 22 4 Old Sources o 0 9 0 RingdSources4 30, Defueled, r'.

~w9,3.: 1 j

o-U-

, EE-DEC-0022 .

l REV B ,i Page-38 of 39- l 1

VI.) CONCLUSIONS / REC 0fEENDATIONS A 1/2 core 3-D diffusion theory model of the FSV core has been developed-to predict startup channel detector response during' defueling. Benchmark comparisons to available experimental data-

. confirmed the model's validity and usefulness in predicting j detector responte for ' core configurations likely to- be-  ;

- encountered during defueling. l 3~ Application of the model for a variety of core conditions during

- refueling Lresults in the following reconmendations for ,

maintaining sufficient startup detector counts: 1 a) . Order two 4.0E+9 neutrons /sec Cf-252. sources for insertion into the core prior to defueling tne last two regions of '

the outer ring of fuel.

i b) Locate the sources in the top layer of fuel in Region 3 Column 3 and Region 6, Column 6.

I c) Use existing - pure graphite reflector blocks-to hold each of the sources, i

L d). Replace the boronated upper: plenum elements in Regions 3, 6.- 10, and 16 with non-boronated materials no later than .

at the time the new sources are installed. .

L e) Defuel the c".;e by ring, starting at the outside (Ring 4)

. and working towards the center.

f) Make Regions 10 and 16 the last regions from Ring 3 to be defueled, l

' g) Continue to monitor startup channel performance to ensure that at least 30 counts /second i s maintained until the m et of defueling.

l i

l {

jt i r

, -w.< w . . - - e , ,--w .o-oo-. *w-

~ ~ ~ "'^ ~ ' ~ ~ ~ ~

} ,. '.-

r -fe ;

s- ,

- . '.i EE-DEC-0022

'l ~

REV B i Page 39 of 39

'i I;

VII.) REFERENCES

1) Malakhof, V., " External Neutron Sourcer. for the Fort St.

Vrain, Unit No.1 Core," GA Document 012472,: Feb. 8; 1973.  :

, 2)' -Sherman, R., " Fort St.- Vrain 3-D Fuel-Accountability i

. Methods and Documentation for Using the FAN 3D System of Codes," PSC EE-ALL-Oll, Sept. 30, 1988.

3) PPS-89-1207. -

4) GP-3296, Merch 23,1989. [

r 7

-i, t

,k

+

l 1 1

.1 l?1 f f i

).

i l -; , , , , -