ML20065U327

Improved Matl Accounting for Pu Processing Facilities & U-235 HTGR Fuel Fabrication Facility

ML20065U327
Person / Time
Site:	Clinch River
Issue date:	10/31/1975
From:	Jeanne Johnston, Mcsweeney T, Schneider R Battelle Memorial Institute, PACIFIC NORTHWEST NATION
To:
Shared Package
ML20065U291	List: ML20065U297 ML20065U312 ML20065U318 ML20065U324 ML20065U327
References
BNWL-2098, UC-16, NUDOCS 8211040167
Download: ML20065U327 (169)

Text

_ _ _ _ _ _ _ _ _ _ . _ . _ _ _ _

l 1

l l

i BMWL-2098 , I l UC-16 l

l l I

i t .

l ,.

l IMPROVED HATERIAL ACCOUNTING FOR FLUT0NIUH PROCES5iNG FACILITIES AND A

{ -

, 8"U - HTGR FUEL rl.Br.ICATION FACILITY ([d' - O 'l i 6 by l

T. I. McSween" J. W. Johnston R. A. Schneider l D. P. Granquist 1

prepared for -

the Nuclear Regulatory Comission as part of the Special Ssfeguards Study  :

, , .7. - -. l

--::.a.;._~;.5 . C O '

-;.::'.L .: C

,,, .. t " L. 7. -.~~ O y - T.J.%1--'" -

~ ~ ~ ~'~ ~

BATTELLE  :::".OL Pacific Northwest Laboratories Richland, Washington 99352 October 1975 I n8 h L'!ST..:007:0.'; rr 7;i~; . , ; ., .7 :;,,, , , . . ,.,[e -

8211040167 821018 PDR ADOCK 05000537 A PDR

BMWL-2098 '

UC-16 TABLE OF CONTENTS Pace i

]

1.0 INTRODUCTION

........................... 1 .

i REFERENCES ............................ 3 l 4 I

. 2.0 Su m ARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

I 3.0 IMP 20VED MATERIAL ACCOLMTING ................... 7 3.1 PURPOSE ........................... 7 j 3.2 FEATURES OF MATERIALS ACCOUNTING . . . . . . . . . . . . . . . 7 .

3.2.1 Objectives of Materials Accounting .......... 7 .

3.2.2 Forms of Accounting . . . . . . . . . . . . . . . . . . 8 }

3.2.3 Limitations in Sensitivity and Timeliness . . . . . . . 9 $

l l 3.2.4 Historical Limitations of Physical ,

Inventory Measurements ................ .

14 3.3 INTERACTION WITH OTHER MATERIAL CONTROL AND "

ACCOUNTING HEASURES ..................... 15 4.0 EXPECTED CAPABILITY OF MATERIALS ACCOUNTINC IN FUTURE PLUTCNIUM PROCESSIf!G OPERATIONS . . . . . . . . . . . . . . . . . . 17

5.0 DESCRIPTION

AND EFFECTIVENESS OF POSSIBLE MATERIAL 21 ACCOUNTING IMPROVE!!ENTS . . . . . . . . . . . . . . . . . . . . . . .

. . .... 21 5.1 IWROVED MATERIAL ACCOUNTING IN SEPARATION PLANTS

. . . .... 26 5.2 IMPROVED MATERIAL ACCOUNTING IN NITRATE STORAGE l 5.3 IMPROVED MATERIAL ACCOUNTING IN NITRATE-TO-0XIDE CONVERSION . . . . . . . . . . . . . . . . . . . . . . . . . . 29 5.4 IllPROVED MATERIAL ACCOUNTING IN FUEL FABRICATION . . . . . . . 31 6.0 COSTS OF MATERIAL ACCOUNTING IMPROVEMENTS . . . . . . . . . - . . 34 6.1 COSTS ASSOCIATED WITH IMPROVED MATERIALS ACCOUNTING IN SEPARATIONS PLANTS . . . . . . . . . . . . . . . 36 6.1.1 More Frequent Physical Inventories .......... 36 6.1.2 Costs of Performing a Running Inventory . . . . . . . . 39 6.1.3 Costs Resulting from Improved Measurement Control . . . . . . . . . . . . . . . . . . . . . . . . 40 6.2 COSTS OF IMPROVED MATERIAL ACCOUNTING IN PLUT0NIUM NITRATE STORAGE FACILITIES . . . . . . . . . . . . . 41 6.2.1 Costs of More Frequent Physical Inventories . . . . . . 41 6.2.2 Running Inventory Costs for a Plutonium Nitrate Storaga Facility ..............- 43

~

5&T ^

av - _

= uw *

. J'

=c:

• - - - - J,' *I ,

"h

7 THIS DOCUMENT HAS BEEN REPRODUCED FROM NRC MICROFICHE HOLDINGS BY

~

i THE DOCUMENT MANAGEMENT BRANCH,

, DIVISION OF TECHNICAL INFORMATION AND DOCUMENT CONTROL i

FOR INFORMATION ABOUT MICR0 GRAPHIC SERVICES INCLUDING EQUIPMENT CALL: 28137 i

{ FOR DETAILS ON:

MICROFICHE BLOWBACK SERVICES CALL: 25076 APERTURE CARD BLOWBACK SERVICES CALL: 28076 MICROFICHE OR APERTURE CARD DUPLICATION CALL: 28076 FOR EQUIPMENT REPAIRS CALL: 28397

'i l

• • e 4

8 m e

T

, , . . . . , , . - . .......~.....-....m-.....,..m..,

m > ,: . --- = . ~ . _ _ _

r_ ._MO i.I r'i c-s, 5

I i4 s. *

, 'e I

? e,e

+4

1. 4

• 'g.

=

H..

N-f O

b h h d.

P l h l\b  :; ,

O- . _s; e, l

9 I. -

l -

t e

t g .

O i

b o

1 s i

=

g y e f

g ..,

. > ;c p  ; .

.

• 7=

v. .,ge

- -n. -Q. .

J e

..'s:,,

2.  :. -- m ~

'*2:

^

-e- w .mdf. " ?, a, .;

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

Pane 6.3 COSTS OF IMPROVED PMTERIAL ACCOUNTING IN l PLUTONIUM NITRATE-TO-0XIDE C0 VERSION FACILITIES . . . . . . . 44 1 6.3.1 Costs of More Frequent Physical Inventories . . . . . . 45 6.3.2 Costs of Inventories Obtained by Process Draindown . . . . . . . . . . . . . . . . . . . . . . . 46 6.3.3 Costs of Running Inventories . . . . . . . . . . . . . 46 6.3.4 Ccsts Resulting from Improved Measurements l Control . . . . . . . . . . . . . . . . . . . . . . . . 47 6.4 COSTS OF MATERIAL ACCOUNTING IMPROVEMENTS IN A 200 MT/ YEAR MIXED C"IPE, FUEL. FABRICATION FACILITY . . . . . . 48 6.4.1 Costs of Obtaining More Frecluent Physica!

Inventories . . . . . . . . . . . . . . . . . . . . . . 48 6.4.2 Costs of Frequent Inuntories of Highly Attractive Material Forms . . . . . . . . . . . . . . . 50 6.4.3 Costs of Obtaining A Running Inventory . . . . . . . . 52 6.4.4 Costs Resulting from Improved Measuremnt Control . . . . . . . . . . . . . . . . . . . . . . . . 52 7.0 NONECONOMIC ACCEP1 ANCE FACTORS OF t%TERIAL ACCOU*iTING IMPROVEMENTS ........................... 53 7.1 SOCIAL IMPLICATIONS ..................... 53 7.2 ENVIRONMENTAL IMPLICATIONS . . . . . . . . . . . . . . . . . . 53 7.3 INSTITUTIONAL IMPLICATIONS . . . . . . . . . . . . . . . . . . 53 7.4 LEGAL IMPLICATIONS . . . . . . . . . . . . . . . . . . . . . . 54

8.0 CONCLUSION

S AND RECOMMENDATIONS . . . . . . . . . . . . . . . . . . 55 8.1 BENEFIT-COST EVALUATION ................... 55 i 8.2 REC 0 mENDATIONS FOR IMPROVEMENTS IN MATERIAL ACCOUNTING . . . . . . . . ... . . . . . . . . . . . . . . . 57 8.3 RECOMMENDATIONS F0P FUTURE RESEARCH . . . . . . . . . . . . . 62 REFERENCES ............................ 64 APPENDIX A - MATERIAL ACCOUNTING METHODS . . . . . . . . . . . . . . . . 65 A.1 THE MUF CONCEPT ....................... 65 l

A.2 STATISTICAL MODELS FOR EVALUATING MUF

. SIGNIFICANCE . . . . . . . . . . . . . . . . . . . . . . . . . 68 I

~

A.3 MATHEMATICAL MODELLING OF FLOW AND INVENTORY 1 MEASUREMENT ERRORS . . . . . . . . . . . . . . . . . . . . . . 73 83 ' . l A.4

SUMMARY

OF NUCLEAR MATERIAL ACCOUNTING EQUATIONS . . . . . . .

REFERENCES ............................ 85 1

W

e l

. , 1

-lii- '

Page APPENDIX B - SAFEGUARDS MTERIAL ACCOUNTING CAPABILITIES OF FUTURE PLUT0NIUM PROCESSING FACILITIES . . . . . . . . . . 86 9

B.1 MATERIAL ACCOUNTING CAPABILITY OF THE SEPARATIONS FACILITY OF A 1500 MT/YR LWR REPROCESSING PLANT ....... 87 8.!.1 Description of Important Material Accounting Characteristics of a Separations Facility ....................... 90 B.1.2 Measurement Uncertainties for the Separations Facility Plutonium Account-ability Heasurements .................

92 B.1.3 Capability of Material Balance Accounting Systems for Plutonium in a 1500 MT/Yr Separations Facility ................. 96 B.2 MATERIAL ACCOUNTING MODELS FOR THE PLUT0NIUM NITRATE STORAGE FACILITY . . . . . . . . . . . . . . . . . . . 101 B.2.1 Description of Important Accounting Features of a Plutonit:n Nitrate Storage Facility . . . . . . . . 104 B. 2. 2 Measurement Uncertainty Estimates for the Plutonium Nitrate Storage Facility .......... 104 8.2.3 Capability of Material Balance Accounting Systems for a Plutonium Storage Facility ....... 105 i

B.3 MATERIAL ACCOUNTING CAPABILITY OF THE PLtJTONIUM NITRATE-TO-0XIDE CONVERSION FACILITY IN A 1500 MT/YR REFR0 CESSING PLANT . . . . . . . . . . . ... . . . . . . 111 B.3.1 Description of Important Material Account-ing Design Features Except for the In ar.d Oatput Batch Operations . . . . .put ......... 114

':. 3. 2 Measurement Uncertainties for the Plutonium Conversion Facility Account 3bility Measurements

..................... 11S B.3.3 Capability of Material Balanw Accounting System for a Plutonium Nitrate.to-Oxide Conve rsi on .'a ci l i ty . . . . . . . . . . . . . . . . . . 121 q B.4 CAPABILITY OF MATERIAL BALANCE ACCOUNTING SYSTEMS FOR PLUTONIUM IN A 200 MT/ YEAR MIXED OXIDE LWR FUEL FABRICATION PLANT

. . . . . . . . . . . . . . . . . . . 125

. B.4.1 Description of Important Material Account-ing Design Features . . . . . . . . ......... 128 B.4.2 Measurement Uncertainties for Mixed Oxide Fabrication Plant Plutonium Accountability Measurements

..................... 130 B.4.3 Capability of Itaterial Balance Accounting Systems for Plutonium in a 200 MT/ Year Mixea 0xide Fuel Fabrication Facility . . . . . . . . . . . . 132 REFERENCES ................. '"

E mi-+.

, =n,- }O $

xmn ==,.~G.

V . _______. a .,_ -.- - - - -- - - - - - '

. . . - -- --. -a - . .~ n vn -, ~~

f 4 -le-Page APPENDIX C - HTGR FUEL FABRICATION PLANT MODEL . . . . . . . . . . . . . 143 C.1 HTGR FUEL ELEMENT DESCRIPTIONS . . . . . . . . . . . . . . . . 143 C. 2 MATERIAL BALANCE MODEL FOR AN P' ' FUEL FABRICA-TION PLANT . . . . . . . . . . .............. 150 C.3 ESTIMATED MEASUREMENT UNCERTAINTIES FOR AN HTGR FUEL FABRICATION PLANT . . . . . . . . . . . . . . . . . . . 152 C. 4 MATERIAL ACCOUNTING CAPABILITY OF A U HTGR FUELS PLANT

......................... 156 REFERENCES

............................ 159

\

l t

e

. . . . . . . . . ........,...-....-.....,mm.--

er,

,. -m- .mwss cu, . q/iT@w w;,.o.. ama:wms;w-wiseur-ca -=wunne:m f ';

_ g. '

TABLES W.g

. . , r Pace

, L .1 Material Balance Accounting Caprbility of Current Processes

'.1

............................ 20 LEMUF Sensitivity to Inventory Frequency for the Separations Part of a 5 HT/ Day Reprocessing Plant ........ 23 i.2 Reduction of Relative Cumulative LEMUF with Time for the Separations Part of a 5 MT/ Day Reprocessing Plant ...... 25

.3 Typical LEMUF Values for the Plutonium Nitrate Storage Facility in a 5 HT/ Day Reproce sing Plant ............ 28 s.4 LEMUF Sensitivity to Inventory Level and Inventory .

Frequency for a Plutonium Nitrate-to-Oxide Conversion Facility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

.5 i

Reduction of Relative Cumulative LEMUF with Time for the Plutonium Nitrate-to-Oxide conversion facility . . . . . . . . 31 7.6 LEMUF Sensitivity to Inventory Level and Inventory Frequency for 4 200 MT/ Year Mixed Oxide Fabrication Plant

3. 7

.............................. 32 Reduction of Relative Cumulative LEMUF with Time for a 200 MT/ Year hixed Oxide Fabrication Plant ~ ........ .. 33 5.o Sensitivity of Loss Detecticn in Pu02 Storage Area of 3 200 f4/ Year Mixed 0xide Fabrication Plant ........... 33 6.1 Sunnary of Lost Production and Additional Measurement and Labor Costs for Haterial Accounting ............. 35 6.2 Maximum Incremental Costs of Inventory Frequencies l

i Shorter than 6 Months . . . . .

6.3

................. 38 Additional Costs Incurred for Taking Running Inventories in a Separations Plant . . . . .

6.4

................. 40 Estimated Annual Costs for Perfurning Physical Irventories in a Plutonium Nitrate 3torage Facility ....... 43 6.5 Estimated Annual Costs for Performing a Running Inventory in a Plutonium Nitrate Storage Facility ........ 44 6.6 Costs of Supplementing Formal Material Balances with F

Running Inventories in the Oxide Conversion Facility . . . . . . . 47 6.7 Increnental Costs of Plutenfun Fabrication Plant

- Physical Inventory Heasurements taken More Frequently than Bimonthly . . . . . . . . . . . . . . . . . . . . . . . . . . 51 6.8 Increiaental Costs of Performing Physical Inventory -

Heasurements Pu0 2 more Frequently than Dimonthly over the Feed Storage Area of a Plutonium Fuel Fabrication Plant

.............................. 51 m

ME l55' ,,W +-

- : .u.. m:c ;w. *.E'

= .n ;

gg- 4 ,, - - --#~~

[

-vi-Pane ,

t 3.1 Benefit-Cost Table of Proposed Material Accounting j Improvements for Plutentum Processing Facilities . . . . . . . . . 56 i.1 ftaterial Flow for a Single Campaign at a Purex Processing Plant (Basis: 750 Metric Tons Fuel Input, o 6-Mon th Ope ra tion) . . . . . . . . . . . . . . . . . . . . . . . . 91

.2 Estimates of Random and Systematic Error for Separations Facility Plutonium Accountability lleasurements . . . . . . . . . . 93

. .3 Running Inventory Measurement Uncertainties for the Separations Facility of a 1500 MT/ Year LWR Reprocess-ing Plant ............................ 95

.1 Measurement Uncertainty Evaluation for the Separations Arca of a 1500 MT/ Year LWR Reprocessing Plant .......... 97

.5 Running Inventory Measurement Uncertainty Evaluation for the Separations Area in a 1500 ifT/ Year LWR.

Reprocessing Plant . ....................... 99 L6 Comparison of LEMUF Sensitivity to Freq"ency and the Type of Physical Inventory Performed to Obtain the Material Balance for a 5 MT/ Day Separations Facility . . . . . . 100 4.7 Effect of the Measurement Control Program on the Long-Tern Measurement Errors for Semi-Annual Inventory Periods Over the Separations Area of a 1500 MT/ Year LWR Reprocessing Plant . . . . . . . . . . . . . . . . . . . . . 102 4.8 Ef fect of the Measurement Control Program on the Long-Term Measurement Errors for Quarterly Inventory Periods Over the Separations Area of a 1500 NT/ Year LWR Rep roce s s i ng Pl a n t . . . . . . . . . . . . . . . . . . . . . 103 B.9 Estimates of Random and Systematic Error for the Accountability Measurements of the Plutonium Nitrate S tora ge Fa ci l i ty . . . . . . . . . . . . . . . . . . . . . . . . 106 B.10 Measurement Uncertainty Evaluation for a Static Pl u toni um ni t ra te S torage Fa ci l i ty . . . . . . . . . . . . . . .

107 B.11 Measurerent Uncertainty Evaluation for a Plutonium Nitrate Storage Facility with Internal Mixing . . . . . . . . . 109 B.12 Measurement Uncertcinty Evaluation fo.- the Plutonium Nitrate Storage Area in a 1500 MT/ Year LWR Reprocess-ing Plant . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 A B.13 Flow for a Single Campaign in a Plutonium Conversion Facili ty for a 2-Month Accounting Period . . . . . . . . . . . . 116 B.14 Estimates of Random and Systematic Error for Plutonium Conversion Facility Accountability Measurements . . . . . . . . 117 ,

B. 5 Running Inventory Measurement Uncertainties for the Plutonium Nitrate-to-Oxide Conversion Facility of a 1500 MT/ Year Reprocessing Plant . . . . . . . . . . . . . . . . 120 ~,

8.16 Measurement Uncertainty Evaluation for the Plutonium Nitrate-to-Oxide Lonversion Area in a 1500 MT/ Year LWR . . . . . 122 l

~

W

e

.~.w m .m.re w w ., m ,.--w e v m.w n m m v. m ,,,.,. ,a,

-vii.

Pace 1.17 Running Inventory Measurerent Uncertainty Evaluation for the Plutonium Nitrate-to-0xide Conversion Area in i

i 1500 MT/ Year Reprocessing Plant . . . . . . . . . . . . . . . . 123

.18 Conparison of LEMUF Sensitivity to Frequency and the

Iype of Phys 4a1 Inventory Performed to Obtain the Material Ba.s. ace for a Plutonium Conversion Facility ...... 124 a.19 Effect of the Measurement Control Program on the Long-Tem Measureewnt Erro.s for the Plutonium Nitrate-to-0xide Conversion Area for a 1500 MT/ Year LWR keprocessing Plant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126

.20 Estimates of Random and Systematic Error for Fabrication Plant Plutonic'n Accountability Measurements . . . . . 131 d.21 Inventory Measurenent Uncertainty Evaluation for Material Held Up in Process Equipnent in a 200 MT/ Year Mixed 0xide fuel Fabrication Facility . . . . . . . . . . . . . . 133 3.22 Beginning Plus Ending Inventory Feasurement Uncertainty Evaluation for Well-Measured In. Process Inventorie<

in a 200 MT/ Year Mixed 0xide Fuel Fabrication Facility ..... 134 B.23 Measurement Uncertainty Evaluation for a 20011T/ Year Mixed Oxide fuel Fabrication Facility . . . . . . . . . . . . . . 135 i.24 Cc:marison of LEMUF Sensitivity tu Frequency aad Arount of Well-Measured in-Process Material included in the Material Balance for a 200 MT/ Year Fuel Fabrication

.alant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136 0.25 Effect of the Measurement Control Program on the Long-Tem Measurement Errors for a 200 MT/ Year Mixed Oxide fuel Fabricatica Facili ty . . . . . . . . . . . . . . . . . 138 8.26 Measurement Uncertainty Evaluation for the Pu0 2 Power .

Storage Area in a 200 MT/ Year Fuel Fabrication Plant ...... 129 C.1 HTGR TRISO and 8150 Particle Olmensions . . . . . . . . . . . . . 146 C.2 Density of l'.ater'als Used in BISO & TRIS 0 Parte les . . . . . . . 148 C.3 BISO aad TRISO Particle Volume and Weight Fraction Data . . . . . 148 C.4 Descri.alon of Core Co.npositions for a 1160 MWe HTGR If,9 C. 5 frGR Material Balance Month Accounting Period I n i ti a l Co re Camp a i g n . . . . . . . . . . . . . . . . . . . . . . 151 C.6 HTGR Fuel Error Com;.onents

' ................... 153 C.7 Inventory Levels and Estimated Measurement Errors . . . . . . . . 155 C. 8 Sumary of Inventory Uncertainty Calculations . . . . . . . . . . 157 C.9 Sumary of Variance Calculation fe. . 'iTGR 2 35U Fuel Fabrication Plant for a 2-Month / ting Period . . . . . . . . 158 g a..

I

~

1. y. .

1 Yn 456 m_--; ,- nn w l =

g. ~ . --

eiji.

FIGURES Pace

[.

Tanale Power Curve . . . . . .

.................. 72 cer.ceptual Flowsheet Purex Processing Plant . . . . . . . . . . . . 89 a -

>1utonium 0xide Conversion Facility Flow Diagram ......... 112 '

Average Material flow Per 8-Hour Shift for M0 Fuel

- Plant . . . . . . . . . . . . . . . . . . . 2 .

.......... 127

,.. i HfGR Standard fuel Element . .

.................. 144

,TGR Control Fuel Element . . . . . . . . . . . . . . . . . . . . . 145 A Coated Particles . . . . . . . . . . . . . . . . . . . . . . . . 147 i

issile Production Sequen:e Sumary . . . . . . . . . . . . . . . . 154 r

,,_ m ,

_m y _J -

, 1A,. - ^'" '^ ^ U ~

-.w -,-- w ' ^ ' ' ' ' ' ' '

7

,,o. m .--.s~ n- ~ -n ewMv=wSWOMN-, .-- . -wm:rm* nm 7

4 1

'l

. .s

1.0 INTRODUCTION

bf

.4 i s This report was prepared as part of the Special Safeguards '

tudy.III The major obje.-tive of the study was to: " provide a ~

, e cematic assessment of the safeguards measures identified in the ,

eaft Generic Environmental Statement on Mixed 0xide (GESMO) and

'evelop a safeguards plan for protection of plants and materials in the plutonium recycle and high-enriched uranium cycle." Studies I

8 rere initiated to investigate potential gains frem the application ,

t f new technologies such as real-time material control. Simulta-j cously, studies were initiatea to evaluate the potential gains '

fron carrying out existing safeguards measures. This study, titled

' Improved Ha terial Accounting," deals with the latter.

Presently, all facilities licensed to process significant

.juantities of special nuclear materials must establish and maintain a system of nuclear materials control and accountability. As part of this system, limits of error based on the SNM measurement un-certainties must be established and, to within the limit of these uncertainties, all material must be accounted for.(2) This report -

describes the safeguards characteristics of a material accounting systen and then describes the material accounting characteristics of future plutonium and a s sU - HTGR fuel fabrication facilities.

There are always some dangers associated with any r;temit to a model a facility which has yet to be built. It is impossible to project the influence of future technological developments.

~

However, useful areas of future development can only be identified by applying present technology to future plants. If these areas Mw __ _ _ _

_......n

2 I

are actually subject to future study and technological developments, l the future facilities will have more favorable characteristics '

I than those shown in this report.

The models shown in this report are not based on the operating i ~

haracteristics of particular plants. However, when possible, data l used in the models are referenced to the performance characteristics

! ~

of material accounting equipment.

This relevant experience is i

documented, thereby providing a basis for future study and improvement.

e i

W 6

i 2

w. . ?s e= _- .m- -

4 g. e W4,.- -e += ***'*e*

.r-

~

. 257

~;=

7bD '

. N'

--.. . gr ,  ;

_r; =, = :

+

, c,m'0.: .

.m;. w 117.:;' _;.

p> mm.

7 r

, _ . - . ~ -

• ~ts um ri u romTi:96ftsym**ro.r-m sewons:mtuww r , - .:% - --.--

^'

i t

^

REFERENCES I

1) Office of Special St'udies, Special Safeguards Study - Scopes of Work, NUREG-75/060, US Nuclear Regulatory Commission, Washington, D. C., June 1975.

2) Code of Federal Regulations, Title 10 Part 70.58.

i 1

l

f. '

l l

l 1

e l

1 ~.

J 2.0

SUMMARY

This study described the general safeguards characteristics of a material measurement system and then describes typical

! accounting systems for a fuel reprocessing plant, a plutonium ,

. nitrate to oxide conversion facility, a LWR mixed oxide fuel l

fabrication facility, and a high-enriched uranium HTGR fuels plant. .

l These facilities are commercial size, envisioned to be operating in the mid-1980s.

The measurement characteristics of the commercial size plants '

1 differ significantly from presently operating facilities. Present facilities tend to have large static inventories of material whereas future facilities are best characterized as having a high material throughput. Thus, many of the measurement problems 1

associated with present facilities will be less significant in t'uture facilities. Based on the models shown in this report, systematic errors will pose the greatest limit on measurement systems in future facilities. As a result, it is recommended l

that:

• the existing licensing review process for natorial l accounting performance be initiated at the cor.ceptual design stage and follow the progress of the equipment measurement system performance tests through startup; a the present capability of measurement systems, particularly -

with regard to reducing the limit of error of calibrations.

be throughly reviewed and the results of such studies be ,

documented and available in the open literature;

~a

..s T.

((' ' M iiE"

"%e-rr. n ::;~m, 7. ;v .

,,- ~ ~ - .~- ~~

~.  ; _,

v

, . . . ~-..r ~.e,- m e m _ _ _ _ - m_____ _ _ .

~1 i

l

.i U

e regulations governing cumulative LEMUF and series of MUFs "

i be developed to supplement current LEMUF regulation for ill single accounting periods. .

, , resent regulations require a measurement control program as part f f an accounting system. When used with the above reccmmendations, f i

. Me material accounting characteristics of future facilities should ,

e greatly improved. In addition to these genersi findings, i }

l i t

1provecents specific to a particular facility were also identified.

l  :

(I i

I For the chemical separations plant, it is recommended that:

i i . formal material accounting periods be ro less frequent .I l

than quarterly for the separations area;

. studies be initiated to demonstrate the accuracy and use-f fulness of " running inventory" measurements;

. formal material accounting periods be no less frequent '

than monthly for the plutonium nitrate storage area; e daily weight factor (specific gravity times liquid height) readings to be taken on static plutonibm nitrate storage tanks to ensure that no change has taken place.

For the plutonium oxide conversion facility, it is recommended that: .

. formal material acccunting periods be no less frequent -

than monthly;  ;

e

. formal raonthly inventory measurements be supplemented 2 with informal inventory measurements whenever a process

~

runout occurs.

2 For the mixed oxide fuel f abrication plant, i t is recommended that:

s t

• ,ade m

_a , 'h

1R P"~

E Y

-e -._ .. - .-.

. formal inventory reporting requirements be no less frequent than monthly;  ;

othe Pu02 storagt area be fornalized as a material balance area and that daily physical inventories using pug 2 weight .

be performed over this area.

The above recommendations, if carefully followed, are envi- .

stoned to greatly improve the material accounting characteristics of future plutonium and high-enriched uranium processing facilities.

In section 8.0, the benefits of an improved material accounting system are cumpared with the costs associated with the improvements.

In most cases, the additional costs range from several thousands to several hundred thousand dollars per year. In one case, the requirement for a formal quarterly inventory over the separations area of a reprocessing plant, an annual lost production cost of

$25M occurs. This is considered to be an extremely high penalty and it prevents reducing the formai inventory requirement from the present semiannual regulatory requirement. In all other cases, the cost-benefit relationship of the improvements is considered

- f avo rabl e.

e e

e l

l .

m _

U" - M tE' i'9?

j.a--g.

Igg g;ggi.

7 F

- .- s. _~s~.=.._

.-- . . _ ._~w=. . . _ _ _ _-

1

• ?

j w

'L 3.0 IMPROVED HATERIAL ACCOUNTING d 3.0 PURPOSE The purpose of the task is to evaluate the s improvements in the sensitivity and timeliness of loss detec- .

tion that realistically can be achieved in currently planned  ;

or future plutonium processing facilities. Improvements in ,

sensitivity and timeliness of loss detection depend upon '

evaluating the safeguards effectiveness of various material i

control practices and equipment design features that could be used to improve the performance of material accounting.

Materials accounting is used in the broad sense as a safeguards function that provides deterrence and detection of diversion and assurance that diversion has not occurred.

l 3.2 FEATURES OF MATERIALS ACCOUNTING l 3.2.1 Objectives of Haterials Accounting ,

The objective of materials accounting I measures is to identify the quantities, locations, and forms .

l l of nuclear material under a particular custodial responsibility.

Ur. der the broad objective, safeguards goals are to detect the absence of material from the. accounting records and to identify the specific area of loss and those immediately responsible r

, for the material.  ;

Materials accounting is the cornerstone of safeguards assurance in that it provides quantitative evidence that the more timely measures for the prevention and detection of diversion have been effective. This quantitative check is l

the af ter-the-f act evidence that the nuclear material is ,

.I

?

l Ia h l J

indeed present and accounted for.

Materials accounting has the virtue of being independently -

verifiable. Th'is feature has particular value to national or international safeguards authorities who can take advantage of their central position in the flow of materials accounting data to obtain independent verification of facility data. For example, advantage can be taken of shipper-receiver data in which two parties both testify to the presence of the stated quantity in the shipment.

3.2.2 Forms of Accounting Two material accounting procedures are employed in nuclear processing facilities: item accounting and material balance accountino. Item accounting, as with the counting of money, is exact and can be repeated as frequently as desired.

The sensitivity of material or mass balance accounting is inherently limited by measurement uncertainties in its power to detect loss. In addition, its timeliness is limited by the need to "close" a material balance, calling for a physical l

inventory to complete the balance.

Item accounting is used as a supplementary assurance measure to mass or quantity accounting to provide visual and immediate evidence that sealed containers, finished fuel rods, etc., are present and accounted for. ,

However, item accounting is not a fundamental assurance measure. The very nature of processing operations require .

that items lose their identity. In addition, the amount of .

material in finished items can be known only through measurements of each item. Thus, the certainty with which all material is  ;

,+, ~

[ "._ 1 2>%r  ??fr )

,. 'W  ?~.2) &W j

7P '

- s s y

n..==- _- - - . - . = - =  :. , - .-

~~

1

. - :t

. y

.J 9- ,

i -1 3i ultimately accounted for is dependent on the..accur cy\of measured , }

.f quantities. ,

Material balance accounts.are taken over a specified period d

of time. The material balance is fo-med by gdding all measur,ed ' "

receipts to the initial measured physical inventory:and t$btracting s i away all measured removals. These additions to and subtrackfEns {$

from the initial measured inventory yield the so-called " book i i

inventory" which is the amount of material that "should be" present. The measured physical inventory establishes what "is" ,

present. The difference between the book inventory and the .

physical inventory is termed " material unaccounted for," ,

abbreviated to "MUT."

3.2.3 Limitations in Sensi tivi ty and Timeliness Mass accounting's value as a safeguards measure is limited by its lack of sensitivity and its lack of timeliness to detect loss. The limitation in sensitivity is inherent in f

that there will always be soma measurement uncertainties.

Materials accounting is also " blind" in that it cannot determine whether a detected loss is due t'o diversion or some other cause.

The Ifnization in timeliness is more practical than inherent.

Tinaliness is limited by the frequency of physical inventory taking. In some instances, a loss could remain undiscovered until the time a physical inventory is taken. There are practical limitations on how often physical inventories can be taken. In many processing operations, there is a real economic need to

~

operate the process on a continuous basis for a reasonable length of time, without shutting down the process in order to take a physical inventory.

~

w

.s,\

5 \s i; ..

s SensIcivity  !

Because of the inherent measurement  :

uncertainty, there is usually some diflerence between what *

"should be" and what "is" - between the book inventory and the physical i nven to ry. Differences also sey develop from unn2asured -

losses and butidup of material in process equipment. The sum of all queh differences is ecpressed as MUF. Since MUF is -

derived frem measurements and subject to measurement uncer-tainties, it may be treated as a statistic. As a statistic s

.MUF has an expected' value and a standard deviation. The s,

expected Value is the true loss or bJildup of material in the process.

The standard deviation reflects whether the s s di f ference

' between what "is* and what "should he" could be due solely to . ".usurement uncertaintias. The standard deviation

• kft{GFassociatedwiththtsemeasurementurcertaintiesis estimated by conbining or propagating the measurement uncer-tainties associated with each quantity entering into the MUF e q u,a ti o n . The standard deviation of MUF arising frcm measure-ment uncertainties is represented by the term aMUF (sigma MUF).

In the Federal regulations dealing with materials accounting,

! the term "linit of error" of MUF, abbreviated to LEMUF. is used rather than a MUF. The limit of error of MUF, or LEMUF, is defined as twice the standard deviation of MUF, e.g.,

LEMUF = 2a ggy.

• The value for LEMUF may be expressed in either absolute or relative units. It is expressed in absolute units in weight .

units of element or isotope, e.g. , kilograms of plutonium.

It is usually expressed in relative units as a percentage of the

• plant throughrut.

Ine statis tical basis for this equation is discussed in Appendix A.  !

Basically.

significance.

LEMUF, defined in this way,is taken to have the following If no unmeasured losses have occurred, then 95 times out of 100, what "should be" and what "it" oresent is ornected ,a .4 4 4 <. . i hA 7, % -

h-* '-i t

a .. 9- gg . 3

'3 r-

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

1 i

A-1 1

1 The power to detect loss is roughly inversely proportional I to the magnitude of LEMUF. The smaller the value of LEMUF '

is, the greater is the power of detection. For a given value of ,

I LEMUF, the probability of detecting loss is roughly directly f proportional to the amount of loss; the greater the loss, the greater the probability of detection.

At this point. several generalizations ebout LEMUF are useful. On the one hand, absolute measurement uncertaintier are always additive. Thus, for a given plant measurement system, the absolute value of LEMUF increases as the amounts of the measured quantities entering into the MUF calculation are increased.

Similarly, for a constant flow process, the absolute value of LEliUF increases with the length of the accounting period simply due to the increase in quantities and their associated un.

I certainties.

By contrast, the relative value of LEMUF when expressed as a percentage of throughput, tends to decrease with increasing throughput. This stems from two reasons, the measurement un-certainties represented by LEMUF have both random and systematic components. The random components decrease in a relative sense with replication or as the number of measured items entering l

l into the HUF calculation increase. The systematic components l

of error, by contrast, tend to remain constant in a relttive sense

~

with replication. The second reason for the decrease in the relative value of LEMUF as throughput increases in that uncer-

~

tainties arising . rom inventory measurements make a smaller

,, contribution to LEMUF.

'i l! .

y -

i 6 '

?

. ~. .. _ -

i It is apparent from the above discussion that the magnitude of LEMUF, which is the index of the sensitivity of loss detection, is dependent on the lengtn of the accountino period. This rela- .

tionship between the sensitivity and timeliness of loss detection is explained more fully in the following discussion. .

Timeliness The timeliness with which a loss (MUF) can be detected is a function of both the frequency of inventory taking and the time required to make the measurements associated l with closinq the material balance. In general, the former is the predominant factor.

In considerinn the relationship betwen frequency of physical inventory taking and detection of loss, several aspects are i

, i r.p o r ta n t . These include:

i

. The promptness with which a diversion could be detected and action taken to recover the material.

. The probability with which a loss " event" could be detected.

The probability with which a rate of loss could be I

detected.

A loss event is defined as 3 relatively large loss that occurs in a very short period of time and during a single ,

accounting period. A rate of loss is defined as a relatively

~.

small loss (or buildup) that takes place continuously over a nunber of accounting periods. . ]

t From the standpoint of providing time for recovery and ' )

apprehension, frequent physical inventories are always of 1

iI advantage. Obviously, the sooner a loss is detected, the sooner i

Me x

=

212: .w:

.. L' 2r 52

~;  :::= + - , =,:

m

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

. , _ _ _ , , , , , __ _m 1

i action can be taken to recover the ria teria l .

The probability that a loss event can be detected deper.cs on the amount of loss and absolute value of LEMUF. As was noted earlier, the absolute limit of error can be reduced, thereby increasing the probability of detection in two basic ways:

first, by improving the quali ty of measurements and secondly, by reducing the length of the accounting period, i.e., incre1 sing j t-the frequency of physical inventory takino. This latter approach, however, has practical and technical limitations.

Physical inventory takinq can incur additional coste because of lost production and additional measurements. From a technical standpoint, a point of diminishing return is reached where the taking of more frcQuent physical inventories reduces the limit of error only slightly. This is the point where most of the contribution to LEMUF is due to inventory neasurement uncer-tainty, with the uncertainty due to measurement of receipts and removals playing only a small part.

D,* contrast, the sensitivity with which a " rate of loss" can be detected generally increases with time. A small but constant loss or buildup occurri.J over a number of matarial balance periods is best detected by evaluation of the :umulative l* HUF. In this case, the contribution of the inventory error '

l plays a less important role since only the beginning ir;ventory of 1 .

the first period and the ending inventrey of the last period contribute to the measurement uncertainty of cumulative MUF. , ,

j Since a constant or persistent loss accumulates 1, time, a point ,

is reached where the ratio of the accumulated los; to the accumulated l

measurement uncertainty is optimum from a cetection stardaoint. u

k.

5 i

i t

• l In summary, the best sensitivity for detecting a large ,  !

single thef t or loss event is obtained by taking frequent physical inventories. Whereas the best sensitivity for detecting a ,

small continuous thef t or a continuous loss occurring over a number of accounting periods is by the evaluation of cumulative .-

)

MUF. In detecting a loss event, the observed MUF of an individual accounting period is evaluated against the measurenent uncer-tainty of MUF for that period. For detecting a rate cf loss, the cumulative MUF for a number of accounting periods is evaluated against the cumulative measurement uncertainty.

3.2.4 Historical Limitations of Physical Inventory Measurements In certain process designs, the assurance that could be obtained from short accounting periods was limited by the exactness with which the in-process inventory could be

measured. Of ten, large quantities of inventory were present at locations in which the quantities could be only crudely estimated or were in forms inherently difficult to measure. As a result of these difficulties, research and development efforts were initiated in recent years to provide equipment design guidelines <

1 for reducing those limitations. The ef forts have resulted in design l guidelines II-4) aimed at achieving the following objectives:

. The transfer of the bulk of the in-process inventory to locations wher( it can be measured with a high degree . l of exactness.

. The incorporation in the process design of systems for

• 1 measuring the material in process vessels.

l

. The incorporation of process steps such as in-house scrap l

J

=- .

] 4%+ WM-mr ano 5.*~ ~ $$E , %d. '

? r-

. . . . . - - = - _ - .

. . . , - - = .-ar avow =- . -. - = a=.

?

q recovery that convert dif ficult to measure materials to  ;(

1 more readily measurable forms. -

Currently designed processes are expected to include features that allow mass accounting to reach measurement sensitivities of loss detection approaching the state of the art.  !

3.3 INTERACTION WITil 0 tiler f1ATERI AL CONTROL AND ACCOUNTING 11EASURES Often the ir.herent limitations of mass balance c-accounting are assumed to apply to an entire facility. This tends to create the false impression that (1) a loss can be detected only af ter a physical inventory has been taken and (2) the sensitivity of detecting a loss is always limited by measure-ment uncertainty.

Normally, mass balance accounting is applied only to the processing activities. Item accountinc is usually applied to materials held in storage awaiting processing, shipping, etc.

This form of accounting can be very timely. If material can be properly sealed and contained, seals can be checked without measurement error.

Formal materials accounting is usually applied to a whole plant or over several process steps. Usually, the best state-1 of-art measurements are used for accountability purposes. These exact measurements are often fairly expensive and often require ,

e several days of analysis time. Less expensive measurements with shorter turnaround times are normally used for process control purposes. Because of the immeoiate and local nature of process control activities, most large loss events would be discovered

~

quite promptly and the area of loss quickly identified. The staff expects that the RETIMAC concept, supplementing process Y

l l

9P

~ ~ -

control data with NDA instrumentation to obtain "Real Time Material Control." will provide eventually a major improvement in that form of thef t detection.

As part of the evaluation of mass balance accounting, the possibility that individual process steps will be able to provide ,

much faster response times than the overall process balance will be considered. If a weight balance is maintained around each process hood or separate step, even though it is not a part of the official accountability system, it does provide immediate assurance that materials have not been simply taken from the process area.

For some process steps, daily records of process yields are

, maintained. These can also provide a quick Indication of the direct removal of material.

e G

O e

L Y~ 2 .E

~

[.h chh. hhh-

- .= ,

'.1 d

}s 4.0 EXPECTED CAPABILITY OF MATEPIALS ACCOUNTING ~

I N FUTURE PLUT0N I Hit PROCl SS I Nil GPL'iiXT IONS  :;

-t, As described in Section 3.2, the estimation of ~y .

u. ,_kl _ &&

?.:

• , Y ~

== x 3;; :gy ==~

.f

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

. =

• t 5

3

'1 The last three columns are the most important. Column S  ?

a shows the absolute value of LEMUF in kilograms. Column 6 -

N shows the relative values of LEMUF as a percentage of two forms of plant throughput - (1) feed plus inventory and (2) i I

feed only. Column 7 shows current Regulatory requirements for relative LEMUF.

l The values for LEMUF which are shown in Table 4.1 are based

• an state-of-the-art measurement quality and current Regulatory -

requirements for frequency of physical inventory taking. The detailed material balance models, the estimates used for random and systematic error of measurement, and the error propagation models are described in Appendixes A and B. .

The values presented in Table 5.1 are for plants which are presently in the design stage or are being built. State of l

the art measurement techniques were used in propanatine measure-ment uncertainties. Thus, except for HTGR f abrication plants, all plants exceed present regulatory requirements. The following sections will evaluate additional improvements in the effective-ness of material balance accounting which might be realized in future plants.

~.

{

l TA3LE 4.1 Material Balance Accounting Capability of Current Processes i.

l I

el #2 43 '

1. 4 -

1. 5 -

a6 ~

1. 7 Special Inventory Accounting Nuclear Feed Level LENUF Present Capability Regulatory

_ Process Period Material kgs kgs kqs 5(Feed +1nv.) IFeed Requirement.%

Separations 6 months Plutonium 7568 54.4 0.72 1.0 5 0.72 Pu(NO3 ) St.stic Plutonium 0 2000 5.16 0.26 -- 0.5 Storage

' Pu(NO3 ) 2 months Plutonium 2500 2000 17.38 0.39 0.70 0.5 Storage Pu Nitrate- 2 months Plutonium 2500 h

1.5 9.59 0.38 0.38 0.5  ?

0xide

, Conversion Plutontun 2 months Plutonium 1190 772 3.67 0.19 0.31 0.5 Fuel Fab.

35 HTGR Fuel (a)2 months U 1632 1520 36.82 0.86 2.25 0.5

, Fab.

(al Shown for comparative purposes only.

lhh

~ , ,

. 4

. O g g

j i

)

)

5.0 DESCRIPTION

AND EffF.CTIVENESS OF POSSIBLE 1 MATERIAL ACCOUNTING IMPROVEMENTS  ;

i The previous section illustrated material balance ',

a

-

• accounting capabilities for currently required inventory b

. 1 frequencies and state-of-the-art measurements. The material  ! l, bali nce areas studied extended over an entire processing operation of the entire plant.

This section considers the possible increases in safeguards effectiveness which can be achieved by improvements in the material balance accounting methods used in plutonium fuel cycle facilities. To increase the safeguards effectiveness of material balance accounting four basic improvements are desired. ,

These are:

. Improvement in the sensitivity of detecting loss events.

. Improvement in the timeliness of loss detection.

. Improvement in the sensitivity of detecting small losses (or buildups) which occur over c long period of time.

. Improvement in the extent to which loss detectior is localized.

! 5.1 IMPROVED MATERIAL ACCOUNTING IN SEPARAT'ON PLANTS

! The reprocessing plant evaluation presented in l

section 4 was based on a 5 MT/ day plant using the Purex process. -

Because this process has been used by both government and

. private industry there is a wealth of experience concerning the plant's operating characteristics. ,

Basically, the plant is a highly integrated facility. This -

integration is required in order to obtain the desired amount of separation of the uranium and plutonium from the waste .

.1 L

94 P- l l

l

)

-zz- .

The plant is designed for continuous operation.

products. .

Industry personnel have stated that it is hiahly desirable to run a reprocessing plant continuously for several months and then shut down for an extended maintenance period. Thus.

plans for the AGNS plant at Barnwell contemplate continuous opera-tion for 5 months followed by a 1-month shutdown period for This mode of l maintenance and cleanout for inventory taking.

)

operacion was used in calculating the LEMUF values shown previously in Table 4.1. f The gains in safeguards effectiveness which can be I

achieved by increasing the frequency of inventory taking are shown in Table 5.1. The absolute values for LEMUF shown in Column 3 of Table 5.1 show the increased sensitivity for detecting a loss event which can be gained by increasino the frequency of inventory taking. The relative values for LEMUF shown in 1 Column 4 reflect the sensitivity for detecting a rate nf loss which is proportional to throughput.

Two plant conditions for inventory taking are shown in Table 5.1. First is a cleanout inventory which entails up to I 2 weeks of flushing and cleanout. Second is a running inventory in which the plant is inventoried without a process shutdown. This second inventory method does, however, involve

. j careful planning and operational adjustments to obtain the best conditions for measurement of the inventory.

2 As the data in Table 5.1 shows, a running inventory yields the same magnitude as those associated LEMUF values of about with a cicanout inventory. This comparison is misleading from The c!canout inventory the standpoint of safeguards assurance.

includes all the plutonium in the process equipment.

nie c

+ g -~- .g mp

. . . -m-5 ;iikh* ..

7F 4

3 i

i

- j i

?

TABLE 5.1- LEMUF Sensitivity to Inventory Frequency for the Separations Part of.a 5 MT/ Day Reprocessing Plant

1. 1 #2 #3 #4 Accounting Inventory LEMUF x 100 Period Feed Kilograms LEMUF Feed Cleanout .

Inventory (kg of Pu) -

I week 303 5 4.09 1.35 2 weeks 605 5 5.85 0.97 l

1 month 1251 5 10.07 0.80 l 2 months 2523 5 18.85 0.75

) 3 months 3784 5 27.8 0.73 i

. 6 months 7568 5 54.4 0.72 12 months 15136.5 5 107.8 0.71 l

Running g,)

Inventory I week 303 222 5.61 1.85 2 weeks 605 222 6.99 1.16 1 month 1251 222 .

10.77 0.86 2 months 2523 222 19.24 0.76 (a) Running inventory includes only the plutonium present in liquid form. It does not include material firmly held on surfaces or soluble forms of plutonium adhering to surfaces above the normal liquid level of equipment.

M

% *um e- e e= + s

1r By contrast, the running inventory includes only the pluto-nium present in process liquids. It does not include material ,

firmly held on surfaces or material adhering to surfaces above the normal liquid level of equipment. Experience nas shown that such accumulations can add up to 10-20 kilograms of plutonium.

As a consequence, there doesn't appear to be any practical substitute to periodic cleanout inventories entailing a fairly thorough cleanout. The running inventory does provide a rapid assurance check and is worth evaluating on a trial basis. How-ever, it should not be used as a substitute for a cleanout inventory.

Improvements in the measurements themselves are expected to further enhance material accounting's value as an assurance measure. Definite improvements in measurement quality are expected to result from implementation of recent regulations calling for formal measurement control programs.I9) One result expected from implementation of such programs is a reduction in the long-term systematic errors of measurement. The effect uf such a reduction is shown in Table 5.2 for cumulative LEMUF(10) taken over a number of accounting periods. The decreasing trend in relative LEMUF with time is due to a partial canceling of f

the systematic error of flVF. That is, some of the systematic l

j errors from one accounting period have dif ferent directions than .

those from the next period and thus are cancelled.

Increasing the ' number of material balance areas within the separations part of a reprocessing was not considered beneficial. ,

The current requirement of separating the plutonium product area from the separations part is considered suf ficient for purposes of localizing losses and providing extra protection for attractive forms of material. -m, f-=

• 7th

=n::

3r

n

'Y$L  !.$$

- _. - ~ . -

l t

b

]

.a 5

TABLE 5.2 Reduction of Relative Cumulative LEMUF with Time -

for the Separations Part of a 5 MT/ Day Reprocessing ,

Plant Length of Accounting Number of Cumulative Period Periods Years LEMUF, % Feed 6 months 1 0.5 0.72 2 1 0.54 4 2 0.37 8 4 0.29 16 8 0.20 3 months 1 0.25 0.73 2 0.50 0.52 t

3 0.75 0.43

" 4 1.0 0.38 l!

5 1.25 0.34 l

6 1.50 0.31 7 1.75 0.29 8 2.0 0.28 12 3.0 0.23 16 4.0 0.21 20 5.0 0.19 l

l b

mW i

i 1

5.2 IMPROVED MATERIAL ACCOUNTING IN NITRA) STORAGE The plutonium nitrate st; rage area of the repro-cessing plant consists of eight banks of slah tanks. These are used foi receiving, storing, and mixing the plutonium nitrate product from the separations part of the process. The storage ,

area receives feed as concentrated plutonium nitrate solutions from the separations MBA and ships plutontun nitrate solutions to the conversion process for converting nitrate to oxide.

The basic materials accounting objective for the nitrate storage area is to provide assurance that the plutoniun credited to the storage account is present and accounted for.

The exactness with which this objective can be net will depend un opera tional s tatus of the various s torage banks a t any particular time.

Severa l opera tional conditions need to he considered.

These are:

. Banks (of tanks) in static conditions where the solution volume (or weight) is essentially constant over the accounting period.

. Banks in which only internal mixino has taken place or f ron which only shipmer.;s (trans fers out) were mace during the accounting period. .

. Banks in which solutions of a different plutonium con-centra tion were received and mixed with previously ..

sto.ad material. -

The timeliness and sensitivity of materials accountino $

assurance checks will be somewhat different for each of the I above conditions, i

^ ~;' '

. * * *e

. .~_. .g .

. rm.  ;; w

. . l e. :L .: 3 -;r

. i ?* ..

~ g;

For banks of tanks in a static condition, nearly con-tinuous assurance is provided through in-place instrumentation which records the weight of solution in each tank (the weicht of solution above the bottom dip-tube). Also, when only internal mixing takes place within a bank, a weight balance can readily be obtained on nearly a daily basis. Similarly, when a bank is used only for shipments, a daily volume balance is also possible.

By contrast, the least t!.nely and least sensitive situatien would be when a bank of tanks is used for receiving and mixing new material on a continuous basis. In this case, a material balance involving both solution weight and plutonium concentration would have to be formed around those i banks. This balance is least timely because an accurate l

! inventory is not possible until the tanks are completely mixed

(

(a several-day operation). It is least sensitive because the 1

physical inventory measurement includes both volume (or weicht) and assay errors.

The timeliness and sensitiv^es of the various kinds of material balances which may be used for safenuards assurance at the plutonium nitrate storage facility are shown in Table 5.3.

, The sensitivities for loss detection are shown by the LEMUF 1

values. As the data show, the smallest LEMUF values arc for .

. material in a static conjition wnereas the largest values correspond to the situajion where material is received and mixed on a continuous basis.

,; The values for LEMUF shown in Table 5.3 should be regarded I

as illustrative of the various operational conditions which

.... ..<.s s. su. .s..... . . . . . .. . . .. .

. ~ -

TABLE 5.3 Typical LEMUF Values for the Plutonium Nitrate .

Storage Facility in a 5 MT/ Day Reprocessing Plant Accounting Ope ra t i ona l Feed inventory LEMUF Period Conditions _ Kilonrams I day Static 0 1000 3.46 I day Static 0 2000 4.89 I day Internal Mixing 0 1000 3.65 1 day Internal Mixing 0 2000 5.16 1 day Shipping-Receiving 50 1000 5.85 and Mixing 1 week Shipping- 300 1000 6.15 Receiving and Mixing 2 weeks Shipping- 600 1000 7.47

, Receiving and Mixing 3-4 weeks Shipping 1000 1000 10.15 Receiving and Mixing 2 months Shipping- 2500 2000 17.38 Receiving and Mixing 1

0 e

i !

h m me' w.r o l,[y l R&&R' ?l5~

I 1*- ' . .

. '}{ - - . 1 30

• • ~f* Ujy [.

, r-1 1

{

s i

i that a mixture of conditions will exist, e.g. , some tanks will 4

be static, some under internal circulation, and some in which  ;

shipping, receiving and mixing are taking place. s 5.3 IMPROVED HATERIAL ACCOUNTING IN NITRATE-TO-0XIDE I~

CONVERSION The sensitivity and timeliness of the ,

material balance performed over the nitrate-to-oxide conversion process using oxalate precipitation was evaluated in a manner which parallels the analyses performed for the other processes.

The resul ts are shown in Table 5.4.

Two levels of inventory holdup are shown in the table.

l One is for a cleanout inventory in which the entire process is cleaned out, including the calciner. The second case is for a draindown inventory which involves a nitric acid flush of all the equipment except the calciner. In the draindown inventory the calciner holdup is reduced to a minimum level, but the furnace is not flushed. The draindown inventory requires about a one-day proces: shutdown and the cleanout inventory requires about four days of process shutdown.

A running inventory was also considered but is not shown in the table. The uncertainties associated with this type of inventory were too large to be of safeguards value.

These large uncertainties stem mainly from the large quantities of plutonium in difficult to measure forms such as highly hydrated precipitates and powders. In addition, there are large quantities present in piping and in the calciner. In these locations amounts can be estimated only crudely. -

u,

. m p-EP_C _ - mt _ __ ,

4

3. .

s

7m h

l 7

1 . - . .

I As the data in Table 5.4 show, both the timeliness of g detection and the sensitivity to detect a " loss event" are -

improved by increasing the frequency of inventory taking. This fi is cor.trasted by the relative values for LEMUF which show that j i

the sensitivity for detecting a rate of loss increases with l l

time. "

TABLE 5.4 LEMUF Sensitivity to Inventory Level and Invectory Frequency for a Plutonium Nitrate-to-Oxide Conversion -

Facility Accounting Feed Inventory LEMUF LEllUF Period kilograms  % of Feed I week 300 1.5 1.798 0.599 1 week 300 4.0 3.119 1.040 2 weeks 600 1.5 2.730 0.455 2 weeks 600 4.0 3.735 0.622 ,

1 month 1225 1.5 4.922 0.401 1 month 1225 4.0 5.543 0.452 2 months 2500 1.5 9.589 0.384 2 months 2500 4.0 9.998 0.400 The improvement in long-term assurance which is expected to result from current requirements for formal measurement control programs is shown by the cumulative LEMUF values given in Table 5.5. The values shown should be interpreted as illus-trative of the expected trend rather than interpreted in an .

absolute sense. ,

e '

?

I i

e - - . . - - --

..u. .. :

~:

n n,

y,g n..

3,ug,.

. :z -

5 3-~j~[

3 5

, . . . ~ . . . . . . . . __. __. - _~ __ . _ _ . =

t-

-al-

.{

3

'i TABLE 5.5 Reduction of Relative Cumulative LEMUF with Time for the Plutonium Nitrate-to-0xide Conversion .

Facility >

Length of 5 Accounting Number of Cumulative ,

Periods Periods _

Yea rs LEMUF, % Feed 2 months 1 0.166 0.384 i 2 months 3 0.5 0.234 l 2 months 6 1.0 0.179 ,

~

2 months 12 2.0 0.144 }

2 months 18 3.0 0.131 -

5.4 IMPROVED MATERIAL ACCOUNTING IN FUEL FABRICATION The timeliness and sensitivity of material balance accounting for the plutonium fabrication facility was evaluated in the same manner as was derived for the other processes. The sensitivity of LEMUF to inventory level and inventory frequency is shown in Table 5.6. _

l As the data show, both the timeliness and sensitivity for detecting a loss event are improved as the frequency of .

inventory taking is increased. As was noted previously, the sensitivity for detecting a rate of loss which is proportional to throughput increases with time.

Two inventory levels are considered in the analysis.

One is for a draindown inventory in which the in-process inventory is reduced to 6 kilograras. The other is for a

, similar draindown inventory of material in difficult to measure locations, but includes over 500 kilograms in weigh tanks

_ _ . . . . = _ . . - - _ _ _ . _ .__ _ _.- =a y , ~ . - n.

- i l

l I

TABLE 5.6 LEMUF Sensitivity to Inventory Level and Inventory Fr31uency for a 200 MT/ Year Mixed Oxide Fabrication Plant Inventory LEMUF LEMUF Ar. counting Feed Kilograms 1 of Feed .

Period 9 -

1.649 1.10 I week 150 772 2.264 1.51 I week 300 9 1.775 0.59 2 weeks 300 772 2.357 0.79 2 weeks 595 9 2.180 0.37 1 month 595 772 2.276 0.45 1 month 1190 9 3.324 0.28 2 months 1190 772 3.886 0.31 2 months This which allow the accurate measuremert,t of the inventory.

illustrates one of the design features being incorporated in future plants which improves materials accounting capability.

As the data in Table 5.6 show, there is only a slight difference in the LEMUF values for the two inventory levels.

The draindown procedures for inventory taking are estimated to require a process shutdown of about 8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> or one full shift.

The improvement in long-term assurance which is expected to result from formal measarement control programs is shown .

in Table 5.7 for cumulative LEMUF. .

To enhance the loss detection of highly attractive materials, ,

a separate material balance for the Pu02 storage area was ,

evaluated. This balance is based on forming a " weight" storage area. In this area all of balance around the Pu02 i the Pu0 2 is stored in " weigh" hoppers so that a continuous

~ ~ ~- . . . . , , , . , g

-z-.~c,_

'7 Z' NO 'fsri i

~/4 ,;p*p:,9 c$$g.

• .. ~

^'

'*. & ' ~~ ~ q zy

pr i

I.

I  !

i j

s 5

TABLE 5.7 Reduction of Relative Cumulative LEMUF with Time for a 200 MT/Yeer Mixed Oxide Fabrication Plant Length of Accounting Number of Cumulative

• Periods Perinds Years LEMUF, 1 Feed 2 months 1 0.166 0.31 -

2 months 3 0.5 0.18 2 months 6 1.0 0.14 2 months 12 2.0 0.12 2 months 18 3.0 0.11 record of Pu0 weight is obtained. The sensitivity of such a 2

material balance to detect losses of Pu0 2 as a loss in weight is shown in Table S.8. As the data show. LEMUF values ranging from 1 to 2.7 kilograms can be obtained for accounting periods ranging from 1 day to 2 montns.

TABLE 5.8 Sensitivity of Loss Detection in Pu0, Storage Area of a 200 MT/ Year Hixed Oxide Fabrication Plant Accounting Feed Inventory LEMUF(a )

Period Kilograms 8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> 7 290 1.0 I day 21 290 1.0 I week 150 290 1.1

. 2 weeks 300 290 1.9 4 weeks 600 290 2.2 8 weeks 1200 290 3.0 '-

(a) LEMUF values are based on a material balance using weight of pug 2 as the only measurement uncertainty. As su ra nce that substitution has not occurred is obtained through regularly scheduled assays for element content.

_ _ _ . . _ _ _ . _ _ _ _ _ . _ _ _ . _ _ _ _ _ . _ _ ___________________a

ye -

1 l

! l

5 .

6.0 COSTS OF HATERI AL ACC.IllNTIfiti I MPI:'lVLitt,rly  !

l Improvement in material accounting can result in lost production and increased operating costs. Lost production costs are difficult to assess. However, if the facility is shut down to take a physical inventory, then it is conservative to attribute all the lost revenue to the additional inventory requirement. If the facility is down for other reasons, a lost production cost is assessed against the inventory measurement only if the inventory measurement excends the length of the outage. In this case the penalty assessed against material accounting is only for the I additional downtime.

The lost production cost assessment, described above, assumes facilities are independent. Facilities do interact. A reactor cannot use plutonium that was not separated or was not converted or was not fabricated into fuel. The assumption will be made that the facilities can be decoupled enough so that one facility is not solely dependent on the other. Care still must be exercised when comparing facility costs if plants are not sized to process the same amount of plutonium. This occurs in this analysis because the 1500 MT/yr reprocessing plant produces approximately 50 kg of plutonium per day, whereas the conversion plant has a peak capacity of 100 kg/ day and a 200 MT/ year mixed oxide fuel .

fabrication plant is expected to require only 21 kg/ day of plutonium. In the cost analysis, the conversion plant will be .

arbitrarily held to the 50 kg/ day plutonium recovery rate .~ rom ,

separations. The fabrication plant size will not be changed I

1

[,h5 y6

. .p,g-

_w

,.~s b,' .

N ' }\hh a:  : h,,

- .=~ -

l 25:3 :g : 3r .:.g~=  ; 3 .-

,, e -

6 l

t

, {

i but it will be noted in any comparisons that about two reference- j size fabricatien plants will be required to use the plutonium separated in the reference-size reprocessing plant.

The costs of additional labor ano analytical measurements are l

l summarized in Table 6.1. The measurement costs ware based on an i evaluation performed by Brouns and Roberts.O I These costs .

TAGLE 6.1 Summary of Lost Production and Additional -

Measurement and Labor Costs for Haterial Accounting Units Costs for Measurement Control Operations

• Operation $/ Measurement Chemical Assay Pu in M0 40 Pu in Pu02 35 Pu in Aqueous Samples 30 Pu in Waste (a counting) 30 l Isotopic Analyses i Pu in MO 125 Pu in Pu02 'O NDA Weighing Standards 2 Process Items 5 Volume Determination 15 Sampling -

Mixing required 20 No mixing 6 Unit Cost of Labor $150/ man day Unit Cost of Lost Production Separation $180/kg of HMa reprocessed Conversion $400/kg of Plutonium Fabrication $250/kg of HM fabricated I a;teavy Metal l

were estimated from information obtained from ERDA contractors and from a private laboratory which performs measurement services ~

for licensees. A $150/ day chargeout rate was used for the cost 9

~ ~

e of a man-day of labor. These are 1975 costs and include all the ,

charges necessary for full cost recovery. The lost production costs for reprocessing and fabrication were based on economic data presented at an AIF-sponsored fuel cycle conference held in Atlanta, Georgia on March 19-21, 1975.I I The Inst production '

$1/gm cost for the conversion facility quoted at the conference was reduced to $0.40/gm to represent the conversion cost attainable by a 50 kg/ day facility.

6.1 COSTS ASSOCIATED WITH IMPROVED MATERIALS ACCOUNTING IN SEPARATIONS PLANTS Three material accounting improvements were considered in Section 5. These were more frequent physical inventories, performing running inventories, and improvements in long-term measurements in long-term measurement quality.

The costs of each of these improvements will be evaluated in the following subsection.

6.1.1 More Frequent Physical Inventories Because of the highly integrated opera-tion of a separations facility, physical inventory measurements more frequent than once in 6 months, are difficult to accomplish without significant production losses.

Using the Barnwell-Plant as a representative facility, it is designed for an annual throughput of 1500 MT./vav. Based on the 6 MT/ day dissolver apacity, the plant must operate -

for 250 days to process 1500 M1/ year. Thus, shutdown periods

~

• which total longer than 115 days would result in lost production.

An average rate of 5 MT/ day over 300 days allows for rework of .

I off-standard material and for minor process upsets. The remaining 65 days allows for two one-month periods for scheduled plant maintenance and a complete physical inventory,

$4: 4

'~9' .@

weg.

t-t

=. .s.z

. .;." =  ; ' :[-

,. 1 3 ._. , ..f ::.

y e-l l

t l So long as there are no lost production costs, the major lk i

direct cost penalty associated with additional physical inven-tories is the labor costs of process operators workinn in the plant. Of the approximately 150 people employed in the plant. '

{

the . job assignments of 50 people may be affected durino the outage. At $150/ day / man for 14 days, the assigned nanoower cost of each outage is $105,000. If I day's oroduction is lost, the revenue from five NT of fuel at $180,000/MT(12) g3 lost. Thus, the lost production penalty is $900,000 per day.

Using the assumptions in the above paragraphs, Table 6.2 has been prepared to summarize the cost penalties associated with nore frequent physical inventories in a separations plant.

A physical inventory was taken during each 30-day outacc as a base Case. Incremental inventory-taking cos ts were charged when the sole reason for the outace was a physical inventory requircaent. This calculational procedure results in a maximum charoe for taking additional inventories. If, as an example, minor naintenance could be performed durinn inventory shutdowns such that the scheduled 30-day naintenance period could be shortened, then the dollar cost (production loss) assessed against inventory taking would be decreased. No such credits were taken for inventory frequencies shorter than semiannual.

The lost production cost must be taken as a direct penalty to the nuclear economy since additional capacity would have to be installed elsewhere. The effect on Barnwell must be ~

4ddressed relative to their revenue. Thus, for quarterly inventory periods the 140 MT of lost annual capacity represents a 9.3 percent decrease in revenna. --

u-

f

' i.

I TABLE 6.2 Maximum Incremental Costs of Inventory Frequencies Shorter than 6 Months Total Annual incremental Incremental i nc remt.ntal Physical Cost of More .

Annual P hys t r.a 1 Production I'9"'"

Loss I"'*" #Y l> Number Maintenance Inventory b Cost Inventories Requirement Requirement Cost

' Inventory Periods / Year _ (Days) Days' imillions) (millions) (militons) 10rs )

0 0 0 65 0 0 2

28 25.2 0.2 25.4

'T 4 65 28 56 56 50.4 0.4 '50.8 6 65 112 101 0.8 102 10 65 112 126 1.0 127 i' 65 140 140 ,

w 12 f 8 i

1 4-a (4) Base case, two planned 33-day outages with one inventory taken for eachAnyoutane.

decrease (b) The time required for a physical inventory is assumed so be 14 days. e I

or increase in this assumption directly af fects these cos't figures.

Note: (1) Total plant income is assumed at (1500) (180,000) = $270 million/yr.

i l

l 4

! !t - 19 ",

3 .4'

=

3 e e a 4

,e-39 6.1.2 Costs of Performing A Runninq Inventory -

The previous analysis looked at the costs of performing a good physical inventory check more frequently than once per 6 months. Flushing and decon-tamination procedures were used both to bring the heldup material to measurement locations and ? S minimize the remaining physical inventory in the facility. This procedure is time '

consuming and costly.

A, was described in Section 4 the runnino inventory can account for nost of the material present in flowing streams. The costs of taking a runnino inventory are associated with the additional measurements, sampling and analyses required. A computer sof tware package and data collection procedure must be developed.

Although the sof tware package is a one-time charge, estimated to cost $40,000 initially, changes in plant operating procedures and improved measurement sys tems will probably make this program obsolete in a few years. Thus, a fixed cost of $20,000/ year will be used to develop and update the analysis routine. This cost represents an average annual cost. In some years, there may be no charges *, in other ye;rs , major cos ts will Le incurred.

Taking a running inventory requires additional data collection, sampling and analyses. For each running inventory,

. six nonroutine samples must be drawn and analyzed. Assuming t

isotopic dilution procedures must be used on all the samples, '

the analysis cost is estimated to be $200 per sample. Process n.

data collection and evaluation would take an estimated twn man-days at a charge rate of $150/ day. Based on these costs,

~ .

9 i

i 1

each running inventory is estimated to cost 51500.

Table 6.3 summarizes these costs as the frequency of .

taking running inventories is increased. In this table, formal draindown inventories are assumed to be taken semi- .

annually. Foriaal draindown inventory measurements supplant the need for taking a runnino inventory. .

TABl.E 6.3 Additional Costs Incurred for Taking Running Inventories in a Separations Plant Number of(' Total Running Shutdown Inventory Inventory / Days Frequency Year Required Total Annual Costs Quarterly 2 0 $ 23,000 Bimonthly 5 0 27,500 Monthly 8 0 32,000 9

Biweekly 18 0 47.000 Weekly 38 0 77.000 Daily 248 0' 392.000 .

(a) Basis: 10 months of production / year ,

40 weeks of production / year 250 days of production / year 6.1.3 Costs Resulting from Improved Measurement Control Measurement control programs have been instituted to improve the accuracy and precision of raaterial ,

accounting measurements. One improver.ent expected from such programs is a reduction in the long-tern systematic errors in .

flow measurements. As described in section 5 reductions in .

long-term systematic errors can be realized through periodic recalibration of equipment and by ins titutinn improved, sta-tistically based, tests of sampling and measurement procedures.

-vnh ..

7 g. ,;,

. z= = yu a;:n-

~ LM ^

~ ~. '

.:f:  %? ^ . Q[;

,p-

.n..-... . - . . . . . . . . . .

- . - - . .. -~

\

The cost of such a program is estimated to be appro. tina tely

$200,000/ year for a plutonium fuel fabrication plant.III) The costs are expected to be in the same range for the separa tions part of an LWR reprocessing plant. Since the program is part of current regulations, there is no incremental cost penalty associated with impruved measurement control.

6.2 COSTS OF IMPROVED MATERIAL ACCOUNTING IN PLUT0NIUM NITRATE Stoa 4GF FACILTTIES Two improvements in material accountability were considered for the plutonium nitrate storage facility.

First, the costs associated with increasing the frequency of physical inventory measurements will be considered. More fre-

~

quent inventory metsurements result in improvnd timeliness to detect loss. Secondly, the costs of calibrating and usina process control data to esti, mate running inventories will be .

considered. .

6.2.1 Cos ts of More Frequent Physical Inventories The problems of material accounting of liquid plutonium nitrate solutions can be minimized by minimizing the amount of material being stored and by attempting to keep most of the inventory s*.alic during an accounting period. Thus, the operating strategy can have a large effect on the accurr.cy of physical inventory measurements.

The operating requirements vary tramendously over the course ~

of a year. In spite of the variability during the year, there will be times when the storage inventory is low,

.2

~

l l

thare wili Se periods when many of the storaoe tanks are static. '

f At these ti.nes, taking a material inventory minimizes measure- l rent cocertiinties. Full advantage of these conditions should i

be t. ten.

From the standpoint of costs, it is important to consider .

how frequentiy these favorable material inventory periods might ,

)

occur. The salts are also minimized over these intervals.  ;

Favor 31le periods are likely to occur quite frequently. I ria ny + a a.

ece likely to be static for a weak; essentially I i

"e tara w:II te sta tic for a year. The optimum inventory t ine intc,vai tcrresponds to the time it takes to empty or tii1 a storage t ink. At Barnwell, which'is expected to be rearasenta tive ar future facilities, the optimum frequency

.0dd there fore be in the range of from 10 to 30 days.

' During this period one bank might be fillinq, another empty-ing, and the r emaining six could be kept static and locked out. I I

T hi:s . ?civ two se:; of measurenents rather than the complete set ct e e t ,4cuid be  ! quired. As the period lenot' tens, more and Every 2 months, as r .o r e b a r <. s c. f tank. must be inventoried.

f our b mis .iight have been used. For longer periods ,

e.u. y as i: i; 1 i a e 1 y U.a t ansfers or withdrawals fren all of them

.:ould he. o c r. u r r e d .

1,. e c o <. t of or.taining an estimate of a physical inventory ir one bank of tants is about $150. The cost breakdown is: $90 or the sie .'eis,ht-factor measurements, $30 for sampling . ,

a r. d 510 for deter ining the plutonium content in the sample.

Thus, cast is incure."1 only if transfers to or from the bank have c...rre; during the cccountino period. Based on the cost per

.. . . . . . . . . : .7 .. ,

4. -. .

L - --- "T 7pc wit w::- =~4- ,

-.~

_-__~_-__ _ - _ _ _ _ _

Y ~ .~.:.

C

]~2 - -_

QQ - _

} .

5F 1

4

} '

i s

-43

• 1 bank it is now possible to estimate the annual inventory cost '!'

j for several inventory periods and then obtain the incremental d cost difference from the standard 2-month base inventory I period. These costs are shown in Table 6.4. -

TABLE 6.4 Estimated Annua' Costs for Performing Physical '

inventories in a Plutonium Nitrate Storage Facility Measurement Incremental Cost Total Cost from -

Inventory Per Inventory Annual Base 2-month Frequency Period Cost Period -

1 day * $ 300 $ 110,000 $ 106.000 '

I week 300 16,000 12,000 2 weeks 300 7,800 4,200 20 days 300 4,500 900 2 months 600 3,600 0 6 months 1200 2,400 -1,200 ,

1 year 1200 1,200 -2,400 a) Because the storage Facility is a buffer betwe2n the separations ano conversion f acilities, additions or withdrawals may occur 365 days / year.

b) Based on 15,1 MT Pu storage batches / year.

C.2.2 Running Inventory Costs for a Plutonium Nitrate Storage Facili ty Instrumentation to continuously monitor liquid levels in nitrate storage tanks will be required for process control information. These instruments may not be more 1

than one percent accurate without calibration. It is estimated that these instruments could be calibrated so that volume (or '

a . . . . W

, s s en- w -- . ,

weight) could be measured more accurately. Such a calibration is .

not expected to cost more than 51000 annually.

The costs of taking a running inventory are quite small.

taking 2 to 3 man-hours or about $50/ inventory check. .

This inventory check is on the weight factor (liquid level and specific gravity) no samples are taken. As such, it is .

just a supplement to the formal balance discussed previously.

Table 6.5 summarizes the annual costs associated with taking a running inventory.

TABLE 6.5 Estimated Annual Costs for Performing a Runninn Inventory in a Plutonium Nitrate Storage Facility Inventory Total Annual (#}

Frequency Measurement Cost I day $19000 1 week 2900 ,

2 weeks 2000 1 month 1300 (a) A;.Lhies a formal inventory measurement period of 2 months.

6.3 COSTS OF itiPROVED MATERI AL ACCOUNTING IN PLUT0NIUM NITRATE-TO-0X1DE CONVERSION FACILITIES Four improvements in material accountability were considered for the plutonium nitrate-to-oxide conversion facility. First, tne costs associated with increasing the frequency of physical inventory measurenents will be considered. ,

More frequent inventories result in improved timeliness 'to detect loss. Secondly, the costs associated with process draindown will be considered. Third, the costs of calibrating and using

=

m mz n

,mp:

.-  : m

.5 _

,} .

l 1

{

process control data to estimate running inventories will be i I

considered. The final case considers the cost of the measure-l ment control program, f

6.3.1 Costs of More Frequent Physical Inventories The base case reported in section 5 '

performed a formal physical inventory every 2 months. l.

It is estimated that the process would require days to minimize and measure the physical inventory of material '

in the process.

There are two types of incremental costs associated with more frequent inventories. One is the direct cost and the second is the ccst of any lost production. The direct cost is estimated to be $3000/ inventory period. The major cost is the estimated 20 man. days of labor at a charge rate of $150/ day.

At a throughput rate of 50 kg/ day, a 1000 kilegram lot of plutonium from the storage area would be converted to oxide in 20 days. Since the next lot would be expected f.o have different isotopics, normal operations would dictate a rtnout of plutonium I

before the next batch starts. Thus, formal pFysical inventories, taken af ter 20 days of full production may i:apose a minimum stress on f acility opera tions. 'For a reprocessing plant producing 15,000 kilograms of plutonium per year, there could be as many as 15 formal inventory periods each year. The incremental annual cost, at $3,000 per inventory period, would be $27,000. The downtime for the additional nine physical

_ _ , , - -~ " ^

- W.

.r I l

l 1

inventories is 4!- days. The total annual lost production cost for a facility cqnverting 50 kg/ day at a cost of $400/kg is

$900,000. The incremental annual costs associated with a i

monthly inventory period would be $18,000 and $600,000 for additional operating and lost product costs, respectively.

6.3.2 Costs of Inventories Obtained by Process Draindown ,

Rather than a complete cleanout be-tween each batch, a runout of the inventory in the process may be sufficient, in this case the six formal inventory periods could be supplemented by nine draindnwn inventories. The formal inventory measurements are assumed to be taken between batches of plutonium. The costs associated with the nine draindown inventory measurements are estimated to be $600 each 4 man-days of labor. Thus , the incremental costs of drain-down inventories is $5400 annually. The annual lost production cost, based on one shift of production lost / '

Inventory, totals $40,000 annually.

6.3.3 Costs of Running Inventories Running inventories would supplement i

the 15 draindown or cleanout inventories taken during the year.

i The major costs associated with running inventories result from the additional calibrations and analyses required. Since NDA methods are probably the only convenient means for surveying the inventory in process equipment and since the instrumentation *I*

is isotopic dependent, periodic recalibrations are required, {

s probably preceding the conversion of each new 1 MT/ batch of j i

plutonium. j 5

I -

syn j ~~

  ..                                                     -                                mer ~     .r;s 2=: ,   . . . .                             * ~ ~                     "*~*    I   -      -       .

I

         .                                                       The calibration procedure is expected to take 1 man-day of labor and measurements with five process standards for each b

p? of the two parallel process lines. The total calibration cost j is therefore $400 per batch or $6000 per year. The cost j

       . for each running inventory measurement, based on the measurement d
                                                                                                       ^

and labor costs presented in Table 6.6 is $200. It is assumed that f each running inventory measurement will require 10 NDA readings j at $10 each, two volume determinations at $15 each and approx- , imately 4 hours of staff labor at $75. Table 6.6 gives the f. incremental annual costs associated with supplementing formal . inventory measurements with more frequent running inventory ~ measurements. ' table 6.6 Costs of Supplementing Formal Material Dalances with

    }                    Running Inventories in the Oxide Conversion Facility                             '

1

     ?

Running Inventory Incremental Annual Cost Frequency Calibration Inventory Total Monthly $ 3600 $ 2,400 $ G,000 ' Biweekly (25 5000 5,000 10,000 periods /3 ;1r) . Weekly (50 weeks / 6000(a) 10,000 16,000 year) 6000 60,000 66,000 Daily )s"300 days / year a) Based on one calibration every one HT of plutonium processed. 6.3.4 Costs Resulting from Imoroved Measurements l Control The measurement control program has been instituted to improve the accuracy and precision of

                                                                                                    ~

material accounting measures. The measurement control program

  .                                    " - " c;~.__.z_                  ..

,l l ' t j . l-

.m e for the oxide conversion plant would be administered by the same starf responsible for the separations plant control program. Indeed the integration of data obtained from the combined plant inventory measurements would greatly aid the program. Thus, the costs cannot be separated from the $200,000 annual . cost estimate prepared for the separations plant control pro-gram. Because this program is part of current regulations. - there is no incremental cost penalty associated with improved measurement control. 6.4 COSTS OF MATERI AL ACCOUNTING IMPROVEMENTS IN A 200 MT/ YEAR MIXED OXIDE FUEL FABRICATION FACILITY This section will summarize the incremental costs of the four fuel fabrication material accountina improve- , ments described in Section 5.4. The improvements were: more frequent physical inventory measurements, more frequent inventory checks on highly attractive material forms, taking a running inventory and improved measurement quality. 6.4.1 Costs of Obtaining More Frequent Physical Inventories There are periods durinn the course of a year when physical inventory of material in the process is at a mininum. The costs of performing a physical inventory at such times is minimal. Using the proposed Westinghouse Anderson plant as a guide, each of the three input storage silos can hold up to 150 kilo- .' grams of plutonium. At a production rate of 200 MT/ year,150 kg of feed plutonium are used in 7 days of full production. 1

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

m. ... ...- - . + - .

                                                                                                  ..;m . .               -

1 cr--- . , ;,

                                                                        ~             .-          .c.=- -         ..;;;::
           .                  u-         'uy u.=             gy                   ~~.      y          ,

                                         .=   .

e k 0 3 During that time the one silo has been drained, a full one is on standby and the third is being filled.

                                                                                                                -)

The silos containing recovered scrap are cesigned s'imilarly. .. In seven days, one has been emptied, another filled and a third ' on standby. Based on these storage capacities, a formal ) 1 weekly inventory, timed to correspond to the switch from one

   ~                                                                                                               i input tank to another, appears to represent a minimum cost                                      !

condition. In addition, since the completion of one input  ; batch is frequently going to be the signal for a plutonium - isotopic change, runout of the process line is likely to be required. For this analysis, physical inventories taken once l a week are assumed to correspond to both the switch from one input silo to another and the runout of the pellet line. At this time all material in process is weighed and the analytical factor corresponding to its respective Pu0 2 or M0 2 batch is applied to get the plutonium content. In addition, at that time all scrap cans, irrespective of their contents, are sent to scrap recovery. The costs are estimated by determining the number of measure-

• ments required to obtain the physical inventory level, and then t At multiplying by the costs summarized in Table 6.1.
• the end of each period, there could be 58 green pellet boats 135 sintered pellet storage boats and as many as 888 inspected pellet trays in storage.. As many as 18 cans of clean scrap, dirty scrap and waste might be sent to their respective treatmnt e locations. Assuming an average inventory

      .g          of material in the process, about 500 weighings would be required.        The cost, based in the cost estimates in Table 6.1 would                             ,
                                                                                                                    ~

s g 4

!--..-    ,.    -_._f__,,..

i . ~ , _ . . .. ..- ... he 52500. The evaluation of the 18 waste and scrap cans at

               $30 each would cost $500.              Thus the incremental cos t associa ted with taking more frequent physical inventories, at a frequency as high as once a week, is estimated to cost $3000.

Labor costs must also be included. It is estimated s to require one shif t to clean up the process equipment in prepara-tion for each physical inventory. Assuming as many as 20 process operators would be used, the labor cost would be another

               $3000. Thus, the total incremental costs associated with each inventory period is $6000.               This assumes the physical i nventories occur no more frequently than once a week and, therefore, result in no net loss in production.               Table 6.7 summarizes the incremental costs of physical inventory measurements taken more f requently than once every 2 months.

For each physical inventory required the f acility will i experience one shift of lost production. Based on a fabrication cost of $250/kg of heavy metal, the revenue lost by the fabricator l is approximately $45,000. This lost revenue cost must be multiplied by the number of additional physical inventories required annually. f _ These are shown in the last column in Table 6.7. 6.4.2 Costs of Frequent Inventories of Highly Attractive Material Forms The proposed Wes'.inghouse Anderson plant design was used to evaluate the costs associated with , supplementing the plantwide physical inventory measurements with mc w frequent measurements on materials in attractive form. - t The Pu02 p wder input and storage area was evaluat'ed. Unpackaged  : cans under item control can be checked in about an hour, costing i

               $20. Measurement of the weight of material in the storage silos t
                                                    "O"                                      O         'W f,             4F    W9      f           #
        ~
                                                  ^'
                                                                                  ~
                                                                                      = ,     mn        := n i
                                                               . ~:$              : $,        R.       mM                 l

                                                                                                         )
       . . . . . . n.    . . . . . - -       . -
                                                                                                -m 5
                                                            ~51-                                         1 4

costs $15, a calibration check may cost another $15; thus, the ' total cos t of inventoring the plutonium oxide storage area is a i ' approximately $50. 1 Table 6.8 summarizes the incremental cos ts ~ of physical inventories taken more frequently than once every

• 2 months for the plutonium oxide feed storage area . The TABLE 6.7 ~~ Incremental Costs of Plutonium Fabrication Plant Physical Inventory Measurements taken more Frequently than Bimonthly Number of Total Total Lost Inventory Supplemental Frequency Incremental Production Balances Annual Cost Annual Cost Monthly 6 $36,000 $ 270,000 Biweekly 18 108,000 800,000 Weekly 41 246,000 1,800,000 >

TABLE 5.4 Incremental Costs of Performing Physical Inventory Measurements more Frequently than Bimonthly over the Pu0, Feed Storage Area of a Plutonium Fuel Fabrication Plaht Number of Total Inventory Supplemental Increnental Frequency Balances Costs Monthly 6 $ 300 t Biweekly 18 900 Weekly 41 2050 Daily 323 16100 Each Shift 994 49700 l 1 time intervals shown in Table 5.8 are nominal times. The plant is assumed to be operating during each of the 12 months of the year. During the year, 47 Pu0 2 input batches welqhing 170 kg

   ,               will have been processed. The weekly inventory frequency is based on the completion of one input batch per 7 days of operation.

The number of opera ting shif ts is assumed to be 1000 per year. ' Slightly dif ferent assumptions would not significantly af fect the costs shown in Table 6.8.

                                                                                       . g

I

                                                                                                       )

i 6.4.3 Costs of Obtaining a Runninq Invento_ry l Except for the pellet line, essentially all the processes are bat-h opera tion. In additinn, based on the Anderson Plant design, which is thought to be represent.itive - of future plants, a process computer will be used for process evaluation and control. This means that for much of the process

• a running inventory analysis will be available. For the pellet line, because of the holdup in feed and surge tanks, information will not be availab'e except by di f ference. A runout of inventory, while considered possible, would require an average of 3 hours and a maximum of 6 hours. This could not be done any more frequently than once a week, a case which has already been analyzed. Thus, one must turn to real time material control techniques, to get running physical inventory measurements.

6.4.4 Costs Resulting from Improved Measurement Cor trol The measurement control program has been instituted to improve the accuracy _and precision of material accounting measurements. Although t't'is estimated to cost approximately $200,000 III)it brings present plants into com-pliance with present regulations.I9) Thus, there is no incremental cost associated with this accounting improvement. . e

                                                                                   '?E;      --

p.- . 7-.- l

                              ~.       ..         Q:        ;;        . Q..       3&      J55.

5 I k 7.0 NONECONOMIC ACCEPTANCE FfCTORS OF  ! MATLkiAL ACCOUNTING IMPRfvEftt.'i T 7.1 SOCIAL IllPLICATIONS

      ,                                   Employees and the general public accept the requirement that a company he able to account for all the
                                                                                          ~.
      . material in its custody.               Thus, there is total acceptance of improvements in material accountability.

7. 2 ENVIRONMENTAL IMPLICATIONS Based on past history, the best evidence that material losses are occurrino in a process facility has beer obtained from material accounting records. Losses via unexpected pathways were occurring and were not being detected by other means. The existence of unexplained losses initiated an investigation which located the loss path to the environment.

Thus, improvements in naterial accounting will provide assurance that the environmental insult from the facility is belou estab-lished limits.

7. 3 1.iS T I T UT ! 0;l AL IMPLICATIONS Materia! accounting places requirenents on i

operating companies which is unique to the nuclear industry. As such, many companies may hesitate to work in the nuclear industry because of the possibility of bad publicity as a result of poor , . performance. Governmental involvement is high. At the same time, material accounting criteria have been written by both national and inter-national organizations. Thus, improved material' accounting fits l easily into the existing structure of these governing bodies. The impact would therefore be very small. . 1 l i 1

     ,e 1

1 1

                                                                                                                                 ?

i 7.4 LEGAL IMPLICATIO_NS Materf al accounting regulations are presentiv  ! administered by NRC. The legal bases for such a requirement, in the interests of material security, have never been seriously questioned. Any changes to material accounting would have a . minimum impact on the existing structure of regulatory ager.eies, and the way they presently function. s I f i P I

                                                                                           's s
                                                                                       .,,e_m..-                   = , . , -
                                                         .       ~y: -      ev.                 b-             -             -

aw.e . x, =~ .

          , ' ' ' ', . - .      ,~'        -Jr       .;-,       . , , .

y _.-

gr e 4 3 1 4 y _

                                           -GS-8.0   C0flCLUS!0flS A?ID RECn""E'JDATIO?is In sections 4 throunh 7, improvements in material accountability were evaluated both in terms of their benefi ts - improved timeliness and Jensitivi ty - and
   , their costs.       In this section, the benefits and the costs will be compared.        Based on this c,mp.-ison subsequent sections will recommend 1) improvements to present material accounting techniques and 2) will recommend areas of research which appear to have the most potential to furtt.er inprove material accounting tineliness and sensitivity.

3.1 BErlEFIT-COST EVAltlATICT! The benefit of improvements in material accounting are measured as increased sensitivity and timeliness to detect a loss. The costs are neasured by the increased economic burden imposed on the facili ty. Table 8.1 summarizes the estimated benefits of the proposed improvements to materials accountina. Improvements are judged as minor if the change in the sensitivity or timeliness of loss detection is less than a factor cf 2, moderate if they result in changes ranging from 2 to 10, and subr,tantial if they result in changes in loss detection sensitivity which are greater than a factor of 10. This scale is rather subjective, but does judge the relative benefits of suggested material i accounting improve.nents. Based on this scal Table 8.1 ummarizes the benefits and costs of material accounting improvements suggested by the results of earlier sections. The recommendations presented in the next two sections are based on the data summarized in Table 8.1.

i . t' . x , L I i l TACLE 8.1 Benefit-Cost Table of Proposed Material Acc0unting Improvements

   -                                                         a                                                                     for Plutonium Processing Facilities Costs ICI              ~
                !                                                                                                                                                           Benefit            Additional Improved      Improved     Annual              lost
                                                              -                                             proposed Improvement in Material Accounting limeliness                sensitivity OJ eratin3,          Production lj Reprocessing Plant Meterial Accounting

a. Formal Quarterly Inventory Moderate Moderate $200,000 $25 million 77,000 None

b. Weekly Running Inventory Substantial No change Substantlag Moderate 4 None(b) None C. Measurenent Control Program

2) Plutonium Nitrate Storage Area

a. Forma; inventory Moderate Macerate 900 None

b. Daily Inventory of all Static Tanks Substantial Moderate 19.000 None

3. Plutonium Nitrate-Ontde Conversion Facility

4. Formal Monthly Inventory Moderate Minor- 18,000 6 00 ,0') , vi Hoderate 40,000 T

y.' b. Informal MUF Estimate at Moderate Moderate 5,400 Times when Equipment Drained C. Measurement Control Program No change Mo1erate(,) None None

4. Plutonium Fuel Fabrication Facility (d)

a. Formal Monthly Inventory Moderate Minor 36,000 270,033 Substantial Substantial 15.000 Nore l l i .i . J b. Daily Balance over Pu02 Storage Area

   .'O j'                                                                                            c. Measurement Cor. trol Program                   Nrs Change     Moderate         None          None
                                                                                                                             ~

(a) Sensitivitylodetectlossesoccurringat slow rate is improved. (b) Estimated to cost $200,000, presently part of regulations.

   , , . ,                                                                                                  (c) All non-economic costs are instonificant.

Y (d) Approximately 2 fabrication plants are requ red to utillae all the plutonium obtained ~ f I sp from reprocessing. This factor should be included in any comparison.

 .g $k
 .s . :

N g

G{?";r

  ,w I

L i 8.2 RECOMMENDATIONS FOR IMPROVEMENTS IN MATERIAL ACCOUNTING Based on the evaluations presented in this report. the staff makes the following general recommendations: That the existing licensing review process of material accounting performance be initiated at the con eptual-design stage and follow the progress of equipment measurement system performance tests up through startup. luitiating the license review at the conceptual-i design stage can provide the best assurance that those design features which improve the ease and exactness with which nuclear materials can be measured l These

 .                        will be incoroorated into future facilities.

features include the measurement of physical inventories l and material flows. Further, by continuing the review [ process through the startup phase, pilot plant experience can be used to evaluate process holdup and measurement-system performance at a stage where improvements can be made. Lastly, pre-startup calibration data and i measurement tests can be used as a basis for demonstrating l that the system will meet safeguards specifications for materials accounting.

                        .                              That the present capability of measurement systems be the subject of a thorouch review and that the results of such studies be documented and available in the open literature.          At the present time. there is not a complete listing of the present 1

l w...m. . . Y #ees l . e

                                                        .:.                        .  .                               L

 ,e capability of all measurement systems used in material                                                ,

accounting. As a result, designers are not always able to choose the measurement systems which will provide the most acceptable performance. The following recommendations are made for specific facilities:

                    . Chemical Separations:                              The staff recommends that the formal material accounting periods be no less frequent than quarterly.

Quarterly periods give increased assurance that all material processed can be accounted for. In addition, quarterly inventory periods allow for more Thus, sensi-f requent equipment calibration checks. tivity to detect accumulations or losses occurring at a slow rate is much improved. The semiannual inventory requirement simply allows too much material to flow through the system between measurement system calibrations. The staf f_plso recommends that studies be initiated to demonstrate the accuracy of running inventory measurements in a separations facility. Based on the analyses presented here, much of s the inventory in the separations f acility is present in accumulator tanks between processing steps and as . such.can be measured periodically without process shutdown. An actual demonstration is needed because . between inventory periods plutonium has been known to deposit on the walls of process vessels and, as a result, be " lost" urtil the equipment is flushed. In current i

                                       .                       ~ ~ "

• Pr .y ,

                                                                                          -sTi.          ~

6,y. . CN~ ^'

7. ; , . . . . . . . . . . .-- .

[

                                                              - .     =        :. g

 ,e j

designs such " losses" are thought to be small but this 1 must be demonstrateo. Plutonium nitrate storage facilities. The staff recommends the formal material accountino period, be no less frequent than monthly. At any time, plutonium nitrate solution may be held in static or etive tanks. Static tanks are being held for future use and are locked out. Active tanks are those to which material has been added or withdrawn or in which material is being mixed. Because of the impossibility to account for plutonium in incompletely mixed tanks, a running inventory on the facility. is impossible in spite of its simplici ty. Thus, a requirement for a monthly inventory will restrict operations to the extent that tank uniformity must be obtained quite often. At the same time, the monthly reporting requirement makes it highly advantageous to lock out as many tanks as possible for that monthly i period. While it is recognized that the monthly l reporting requirement may somewhat reduce the flexibility of the operations, reduced flexibility is very advan-tageous from the standpoint of material accountability, o Plutonium Nitrate Stcrage. The staff recommends that the static tanks be checked daily to insure that the weight factor (specific gravity times liquid heioht) has not changed, a The static plutonium nitrate storage tanks may be locked out of the process for several accountina

                                                                                                                                      ~

periods. Altho sgh it is not felt that sampling for _ __ _ _ _ . . _ . _ _ - _ .__. - = - - - - ( _ ___ s

                                                     =e=-    .    .[ .

1 plutonium content is required, a daily check of the , weight f actor for each tank would appear to be a minimum requirement. 1 Plutonium Nitrate to 0xide Conversion. The staff ,

              ,                                                                                                                                                            J
                      " recommends that formal inventory reporting require-                                                                                             ,,

ments be no less frequent than monthly. j Dased on the analyses of this process, presented f. earlier, sensitive running inventories of this process are not possible. At the same time, monthly cleanouts are not extremely time consuming and appear to be a reasonable alternative.

                 ,       Plutonium Nitrate to Oxide Conversion. The staff recommends that the formal monthly inventory be sup-plemented with informal inventory measurements wherever a process runout occurs.

Process runouts occur to get a clean separation between batches or to do preventive maintenance. In either case, such runouts provide a convenient time to take an inventory. The results are only slightly less sensitive than the formal inventory.

                   .       Plutonium Fuel Fabrication. The staff recommends that                                                                                      ,

formal inventory reporting requirements be no less frequent than monthly. Future plutonium fabrication plants have very small quantities of material in difficult-to-measure , forms. In addition. process runouts are likely to Thus, providing several occur several times a month. opportunities a month for process inventory measurements.

                      *~                                                                                                                   _ . _, .-
                                                           . _ _ . _ , _ _-   *-  t *    -- - - . .                 . , ,    _m
                                  .m._..     ._m
                                                                                                                                   %                    e,
                                      -             -                          -w                                         . . ;;.y...             -%,..

..L

     . } .Q.f         3

_'[ } .

                                                  . Q: . . *
                                                                             =;;
                                                                            ;:.:.5             Q9. .
                                                                                                                          *m:Q.                 ~

xw.;

 ,w 6

l 1 61-

                                                                                                                                               )

Measurements at such times should be encouraged since

                                                                                                                                               !i they represent what are likely to be ends of accounting                                I frames for other facilities. In this way, cross-checks between facilities may be available.                                                   I
                                              .        Plutonium Oxide                Storage.                                                ;

The staff recommends that i the pug 2 storage area te formalized as a material t balance area and that daily physical inventories { using FuG3 _ weight be performed over this area. Because of the ettractive n sture of this material, which is present in loose form or in sealed containers, the staff believes that daily weight balances and item counts should be performed. Losses from other areas require the extraction of much larger quantities of material, amounts which are likely to be noticed. This is not the case with a Pu0 2 storage area if it is only inventoried with the frequency of the balance of the plant. 0 1

                                                                                                     -      _ ._-                         2_      _,

4

i 8.3 RECOMMENDATIONS FOR FUTURE RESEARCH The following general recommendations for future research' topics is made:

                 . Cumula tive LEM'IF.      The staff recommends that Regula-tions for Cumulative LEMUF be developed for plutonium fuel cycle facilities to supplement current Reaulation for LEMUF for sincie accountina periods.

At the present time, Regulations for measurement quality expressed in terms of LEMUF, apply to only single and f airly short accounting periods. They do not fully address the problem of long-term assurance and the problem. of long-term material control . In some respects, much better assurance can be obtained by evaluating materials accounting data from the standpoint of the cumulative MUF and its associated limit of error, CUMLEMUF, than by evaluating the MUF for a single accounting period. To develop realistic limits for cumula tive MHF, development efforts are required in two areas. First, statistical procedures must be developed for the prop-agation of cumulative measurement errors. Second, studies of current state of the art measurements and their uncertainties must be carried out to provide a . realistic basis for the CUMLEMUF values to be used in the Regulations.

• 1 1

1 I i i. rry a s 93

   .                                                     7                                 75
 ~~                                         .             .r::     :;w          me       - - m:.

sl, L' ,

                               .         'C.'.

L _.f. . [L y

 ,e v>,+-~~ ~ -      n.=:    =_ -n        -..,~._---==<e---

r. e

                                                                                                          ;1
          .                                                                                               a i

4 t 3 ,

                                                                                                          .A       1 Safeguards Assurance of Material Accountino.                                  ::4 '
                        .                                                         The                     3       '

staf f recommends that formal ma terials accounting  : using the best state-of-the art measurements be fostered and improved as a means of providing oositive i

    ,                       assurance that diversion has not occurred.          Further, it is recommended that R&D programs aimed at improving the timeliness and sensitivity of materials accountino be continued and expanded as appropriate.

i 1 e 4 i i r L t 8 - 8 a I l M

                                                             =
                                                                                              ~ k
      '."               '.s                *'.

l_ .~

l 1 k '

                                                                          -b4-                                                                       .!
                                                                                                                                                       )

REFERENCES ,

1. Regulatory Guide 5.8. " Design Considerations for Minioizing .

f Residual Holdup of Special Nuclear Materials in Dryino I and Fluidized Bed Operations " U. S. Nuclear Regulatory Commission, Washington, D. C. May 1974.

2. Regulatory Guide 5.23. "In Situ Assay of Plutonium Residual Holdup." U. 5. Nuclear Regulatory Commission, Washington, D. C. May 1974. .

3. Regulatory Guide 5.25. " Design Considerations for Minimizing Residual Holdup of Special Nuclear Material in Equipnent for Wet Process Operations," U. S. Nuclear Regulatory Commission, Washington, D. C., June 1974

4. Regulatory Guide 5.42. " Design Considerations for Mininizinn Residual Holdup of Special Nuclear Material in Equipment 7

for Dry Process Operations," U. S. Nuclear Regulatory Commission, January 1975. l

5. "Barnwell Nuclear Fuel Plant-Separations Facility Final Safety Analysis Rt ort," Docket 50-332, Allied General l

. Nuclear Services, Barnwell, S. C. 29812, October 1973.

6. "Barnwell Nuclear Fuel Plant-Plutonium Product facility Preliminary Safety Analyses Report," Docket 50-332, Allied l

General Nuclear Services, Barnwell, S. C. 29812, July 1974.

7. " Recycle Fuels Plant - License Application," Westinghouse Nuclear Fuel Division, Anderson, S. C. , July 1973. l

8. " License Application - Gulf Youngsville nuclear Facility,"

Docket No 70-1372, General Atomic Company, San Diego, California, February 1973. l

9. 10 CFR 70.58.

10. K. B. Stewart. " Statistical Techniques for Enhancing the Role of Material Balance Accounting in Safeguards," BNHL-1386, Battelle-Northwest, Richland, Washington 99352, May 1970.

11. R. J. Brouns and F. P. Roberts, "The Incremental Costs and the Benefits of a Proposed Measurement Control Progran for .

SNM Accounting Measurements in a Nuclear Processing ~ ~ Facility " Draf t BNWL Report, August 15, 1975.

12. "Everythings Going Up," Nuclear Industry, Vol 21, #3, March -

1975, p. 13. I I I 9

                                                                                                                                      -q           .
                                                 .7                     7-w Sw                                            -Eg                                ;L.=                           ::,p&
                                                                                                                                    ?

- g  ::. g.; -

m -- _ . 3_ .

e---~,..-. ...,. - . _ __. -,- - m- _ _ M t P

        .                                                                                                                          1
                                                                       -z-                                                       jj
        .                                                           APPLNDIX A                                                     I i

MATERIAL ACCOUNTING METHODS As discussed in section 3, material accountino methods are used to provide assurance that the nuclear material beino processed through a facility can be accounted for. This assurance is accomplished by breaking each facility into a number of discrete material balance areas (MBA's). All material transfers into or out of each .:rea must be measured and recorded. Since all arts of the facility which could contain special nuclear material must be included in a material balance area, a facility balance can always be fonned by combining the results of the individual areas. , In this appendix, the methods employed to evaluate the effects of improved accounting techniques will be described. Appendix 0 applies these techniques, the results of which are summarized in section

3. This appendix will be divided into three sections. This appendix will begin with a general description of the MUF concept. This will be l

l followed by a fairly detailed description of.the statistical models used ( l to obtain the confidence limits on the value of MUF. The final part develops the detaile:: equations used to model possible improvements in { the sensitivfty accounting methods. l A.I Tile MUF C0flCEPT MUF is an acronym for Material Unaccounted For. It is calculated

     ,                 by taking the difference between the book inventory (the material which is supposed to be present in the inventory) and the physical inventory l

(the amount of material which is either measured or estimated to be ~ present). In the absence of measurement, sampling and bookkeeping errors, a positive MUF indicates an unaccounted loss of material and a

                                                                                      .1 N
          ^

4, e , -o "o ' ~ '

                                                                                                   ,     * ^     '*

negative MUF indicates an unaccounted for gain. The book inventory at the end of an accountina period is obtained , by taking the quantity of material present at the becinnina of the accounting period and adding to it all encasured receipts and sub- . tracting all measured removals. Thus, if F denotes the total aniou'It of feed, R all the removals and BI the beginnir.q inventory, then , the ending book inventory, EBI, can be expressed as: EB1 = B1 + F - R. (1) If El is the tr.easured or estimated ending inventory. then MUF is expressed as: MUF = EBI - El = BI - El + F - R. (2) In general, the inventory, feed and removal tems are made up of the sum of many items. MUF can be evaluated for many accounting periods. If in the tth period there are n atches of feed, t removal t g batches, and "k" categories of material on beginning and ending inventory, then: 1 6 E t L t MUF

• IIt-1,1 ~ I t,1) + F t ,1 - R t .j ' (3I t

This expression is for a single material b31ance area. All flows, receipts or removals, from the area are included in the MUF equation. If the MUF's from each area are added up, then the facility MUF is obtained. This occurs because the inventory term for the entire facility eqwals the sum of the inventory in its parts and any transfers between MBA's cancel out. . On the books, all internal transfers are described by ent'ering an R) tem into the records of the MBA shipping material and an identical F, tem in the MBA receiving the material. When the MUF's for these two areas are , summed, since the removals from both areas are subtracted from the feeds from both areas, the tems representino the internal transfer cancel. By definition, feeds or removals from the facility have no such cor-

                                                                                           ;h. --     .w,
                                *           **             : ~*'            v&p          .f & -      4y,
   ^
                                                          ;r
                                                            ~i               5             IiG      3lW
                                                                                                  ,.~.::r = :
                                                                           -c3          33:.

    . , . . . . - . . . . - , . . . - . - - - ~ ~ - ~ ~ ~ - ~ ~ - ~ ~ ~ ~ ~                         -
                                                                                                      -=-   -

M 4 e 3 If the length of the accountino period is very long, then the values ' for ng andi t in equation W can become very lane. De number of categories of material on inventory is not influenced by the length of the accounting frame. In addition it is only the inventory levels at the beginning and end of the accounting frame that enter into MUF. This is easily shown by summing the MUF calculations for two successive periods.

                                                                                                                 -9 K-MUFt + W F    t +1 "

1- It-1 i'I t 1)

• II t.i-I +1.i},

t g (4)

                                                   "t+1                        At+1
                                     +           4 t,i                      tj-                                           '

9 Ft +1,1 ~ j Rt +1.j

• When the two MUF's are sumed the intermediate inventory estimate, I t . i' cancels out. The feed term in the sum of all the feed batches for both periods and the removal term is also the sum of the removals from both i

periods. Thus, in addition to evaluating an entire facility from smaller material balance areas, it ie also possible to combine short material balance oeriods into longer ones. If the MUF's from N consecutive accounting periods are sumed, this cumulative material unaccounted for estimate is designated by the acronym - CU!EUF. When N accounting periods are sumed, equation (3) becomes: . k N

                                                          $                A
  -                                                         t         N CUMMUFg=

(10,k - IN,k) + h f Ft1 ~ j t.f (5) In the above equation the subscript "t" represents the tth accounting period. The tenn 1 0,1 represents a physical inventory term at the

                                                                                                              ~

start of the first accounting period over which MUF's are accumulated. If the cumulative MUF calculation begins with a new, clean facility, then 1 5 **# #* " 0,1 *

                                                                                                               ~

In the following sections. equations (3) and (S) will be used to ,, develop equations for eval'uating the statistical confidence limits that can be piaced on a given value for MUF. A.2 STATISTICAL MODELS FOR EVALUATINt. M.'F SICl!FICANCE The amount of material in a batch can never be measured exactly. Even the amount of material under item accounting is ur.certain to the extent that the arnunt of material contained in each item cannec be known exactly. Placing items under item control does not improve the precision of the inventory measurement; it does simplify and quicken the inventory measurene.it procedures. Recognizing that no measurement can be made without error, the following paragraphs will take estimated errors in basic measurements t and show their effect on the certainty of the MUF and CUf9tVF tems. The basic measures of dispersion and uncertainty in a given measure-ment are the variance and the standard dqviation. These measures of the certainty of a given measurement are described below. The variance of a random variable x is defined by the equation o (x) = (x-p)2f (x)dx. where t: is defined as the mean or expected value of x, and f(x) is the density function and is a measure of the frequency with which x will assume a value in the small interval dx about x. Then o is defined as the standard

                            'eviation.

d If R values are randomly chosen from the distribution, then R x = [x /R g (6A) 1 and 2 I00) - o (x) = (xj -x) /R-1 1 are unbiased estimates of u and 2o respectively. For a normal distribution about two-thirds fall within one standard deviation of the mean and 95

                                                                                                    , T: '    y g
                                                          -              -    + . ,         ..?  +%            M t-
                                                                                                  $ Wh 7^'   'i *.                                  *h -
 . . ~ .[,    ; ,,
           - . . ~ .                .-;~              q.3. .       =p        7.y.'. '     .

4 -.~,-

                                                              - - .. - -- -                    _. ..- ~s t

If two measuremerits are independent, the variance of a series of measurements ecuals the sum of the ir.dividual sneasurement variance. Thus, the equation for the tth accounting period becomes: 2 o (MUF) = c (EIt-1-EI t ) + o F)+ t (Rt }* (} n this equation, the beginning inventory tem fcr the tth, accounting period his been repicced by its equivalent, the ending inventory for the previous accounting ceriod. The F and R terms represent the sum of all

  ,              feed and removal terms for the accounting period. The variance of the inventory tems has not been broken up because the assumption of independerice may not be valid. This is particularly true if the lenoth of the accounting period is very short.

The variar.ce of CUMMUF for N accounting periods can be obtained by a parallel develcpment. The result is : N N 2 2 ohCU*t'fJF) = oj (BIy -El ) + o ( t{ p ) , ,q pt). (8)

 ,                                                          t         t i:

Once again tire variance of the inventory measurement has not been separated because the measurements may not be independent. Equations (7) and (8) can be applied to a single rnaterial talance area of to an entire facility. it should be noted that although MUF for the entire facility is the sum of the MUF's for the individual MUF's, the rela-tionship is not true for the variance of MUF. To be correct, the variance associated with internal transfers must not be included bi the facility calculation. If equations (7) and (8) as applied to NAs were added l together, errors associated with internal transfers would rot cancel but

      /

instead would be counted twice. l l s*-

s - 70 The values of MUF and CUMMUF obtained by closino the ruterial balance can be considered as random variables with a me.m .md a .tae faril eleviat ion. , The central limit theorem would indicate that since MUF .mi (tiMMill are the sum of many distributed variables, where nn individual values tend to be dominating, the values of MUF and CUMMUF will tend to be nomally distributed about the expected valud of MUF (or CUMMUF). The expected value of MUF is the value that MUF would have in the absence of measure- . ment e rrars. If the expected value of MUi' (or CtMMUF) is zero, then the absolute value of MUFg (or CUMMUF) is expected to deviate from zero by less than 2a(HUF) or 2c(CUMMUF) 95 times oat of 100. Because of the statistical sionificance frequently attached to the 95 percent confidence interval, twice the standard dedation of MUF has been given the i.cror.ym LEMUF for Limit of E,rror MUF. LEMUF is usually '4 sed as a control point, i.e., an investigatien is initiated wheneve" MUF exceeds LEMUF for an accounting period. In this way, assurance is gained that all material processed throcch the NBA or the facility has been accounted for. The romal distribution of calculated values of MUF about zero, under controlled condition:, can be used as the basis for several sta-tistical tests. First, if the calculated value of.MUF exceeds LEMUF five times out of iOO no un."easured losses have occurred tJt measurement errors have combined in such a way that the absolute value of MUF deviates frcm zero by a value greater than I.EMUF. When this occurs it is called a

• type I error. Note that if *he expected MUF is zero MUF is ju:t as likely to be more negative than -LEMUF as to oe more positive than +LEMUF. Any , ,

c time MUF exceeds LEMUF an investigation is usually required. or this reason the region outside the interval [-!EMUF,LEMUF3 is defined as the

                                                                           }

critical region. (SeeMood,page247.)

 ^

s-u: , .

   ,~'~s 0- Q '        .  )              _T.         .~~$_'
                                                                             ~

y. e= ~

,' f f QW3

                                                                                                           =

a z. .: . .n w ;

       '~      *   ;; 7

• 3-

e i f I Suppose there is a loss then MUF will be siistributed around t..

                                                                                                                'l c with a standard deviation of oMUF.                  There is sor,e probabt!4ty that                   ;

the loss will not be detected because measurenent errors hide the loss. This is called a type !! error. The probability that a loss c will not

                                                                                                                 ~

be detected can be obtained from the following equation.

 .             p(c) = 1. )r1 IEMUF-..

L

                                                 ,p     -L EMUF-e (11)             ,

dI#MUF ) s "MUF /)* . where F(x) is the cumulative distribution function of the zero mean, unit variance, and norual distribution. Values of f(x) for a given x car be obtaineo from iny standard set of statistical tables.I2I figure (1) shows the value of P(.-) as a furction of r for the case where the expected value of MUF is zero and o(HJF) = 1.0. Equation (11)can be simplified if c>LEMUF. In this case: P(c) = 1 - F ~LEMU c

                                                 =F
                             \ "MUF            )

• MUF /

                                                                        , p ((MUF --2\. /

(12) A similar development can also be carried out for multiple diversions using the expression for en(CUMMUF). If an amount c is lost during each accounting period, then as shown by Stcwart.I I the probability of detecting the cumulative loss Nc is given by: P(Nc) = 1 j F (2cN(ClH4UF)-Nc) /-2c(CUMMUF-Nc)} N

                                                            -F                               II3) j o (CUMMUF)                        "N(CUMMUF)

If Ne > 2a N(CUMhUF) then equation (13) is approximated by: g Nc ) P(Nc) =F1 fNc-2e (CUMMUF))l= -2 F {fL ( "N(cumur) / ( o (CUMMuF) g j (14) t h

• l

i - s ,.

                       ,.                      100 s

2 o $ - =

                ' m' :

o 1 taJ d a u- 60 o P >-

F t-t a

m Sd (Muf) = 1.0 kg .

40 ! il,! 0 4: j. 8 a: CR = 2 Sd (MuF) = 2.0kg a.

                                                              ~

a , l') )," 3 I: I 1 0

                            .                      0                 1            2                           3                   4            5 cjg.                                                      .

'>.L n.>> kg LOST l FIGi1RE A.1 Sanple Power Curve n

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

4 I i

       ,                                                                                                                                                   k' i

Stewart presents the rationale for using equation (12) to detect j single diversions and equation (14) for the detection of multiple L diversions. Equation (14) is insensitive for single losses because the ' variance tems associated with the flows for N accounting periods are - included in the calculation. At the same time, since Nc can be expressed as a fraction of the throughput, the detection probability using c(CUMMUF) increases with time. For short accounting periods, the variance terms ' associated with inventory limit the detection capability. After many accounting periods the variances associated with flow become dominant and unmeasured flow losses become more susceptible tu detection by statistical methods. Bf using equations (7) through (14) it is possible to determine the power curves for a givea set of feed removal end inventory variance reasurements. The following section will look at tt.e detailed error l structure of these tems and describe how basic errors in measurement. l sarer, ling and analysis are cor:bined to obtain the accuracy of a given I flow or inventory measurement. A.3 FMTIIE*tVICAL MODELLIN" 0F F10u A!iD INVENTORY !!EASUREMFHT ERRORS The previous section started with the measurement error variances in I flow and inventory, and propagated these variances to variances in MUF and CUMMUF. Measurement errors associated with various flow and inventory terms are generally statistically independent so that the propagation \

     . technique is fairly straightforward. Obtaining the variance term for                                 '

specific flow and inventory terms is not as straightforward because the assumption of independence is in general not valid. This section will develop the techniques which have been used to obtain the individual flow and inventory variance terms in equations (7) and (8), ' l

                                                                    , . _    . - . . . . . -   - M 3

t There are two factors which make the development of flow and inventory . va-lance terms somewhat complicated. First, the material balance considers a single element or isotope but of ten the measurements are made on mixtures , of elements and isotopes. Thus, a single measurement on each flow or ' inventory component cannot be used. A weight or volume measurement of the entire flow or inventory component must be supplemented by an analysis . of a sample of the material balance component. Thus, the errors in measuring a flow or inventory component must consider errors in analysis, sampling and weight or volume measurements. The second complicating factor is the existence of random, short-term systematic and long-tenn systematic measurement errors. Random errors in measurements are independent errors waereas systematic errors show various degrees of dependency on each other. For example, if a scale is miscalibrated for an entire eccounting period, then every readino taken during the period is in error by a ccnstant amount. Such an error, since it would not t'e expected to persist for many accounting periods, would be classified as a short-tern systematic error. Long-term systematic errors are assumed to persist indefinitely. It is easy to see that when a whole series of measui ments are dependent on a single calibration, the assumption of independence is not valid. Although the two complicating factors are somewhat related, they will l be treated separately. The key assumption will be that it is possible to break each measurement system into random short-term systematic and long-e e , .

                                                                     .         ...       ._   e         w.
                                                   ,J.-                                      .z ,,ci-    ?; r --
     ' L T : ". '    ' *
                                          ^ * *
                                                   "'.T *       ~.'.'        * :; *,." . 3         _
                                                                                                        ^'

e

                                                          -_-.--c.--                                -

i term systematic components. Then all random, short- and long-tem systematic I errors will be treated separately when cr.temining their influence on a l flow or inventory component of the material balance.

                                                                                                             ,i To demonstrate how the individual flow or inventory measurement                      i variance terms are determined, take the feed tem                                            -,
 .                                                                                                           j h

Fg=Wg X, i g (15) .; where F g is the total weight of the element in the ig feed batch, i.e., kg of plutonium. W9 is the total weight of the feed batch (ag), J X g is the concentration of the element over which the material balance is being taken (kg of Pu/kg of feed). Statistical proofs are available to show that an approximation for the variance of the term F g is: 2 " op W X ~ Xg )2 i = (W9

                                        ~

7+gl W X a (16)

                                          ,2
                              = (F )      -f + h,2                    -

(17) W X 2 2 The terms e /W and a /X are the relative error variances associated with weighing and detemining the concentration of the material balance - element or ' stope in the feed batch, respectively. The subscript has been left off the W and X because the relative errors are assumed to be independent of the batch weights and concentrations. Equation (17) will be used to describe the behavior of the random and long- and short-tem systematic errors. For random errors, the ratio a^ - ^ = -

                       ,                                 -a f                                               f                             a  *-

l 1 i i

                                                                                                                                                                                  }

t l , I l of og2/W will be denoted by r2, for short-term systematic errors, the i ratio will be denoted by p , and for long-term systematic errors the ratio l will be denoted by q . The equation for the random error variance of the feed term becomes:

                                                     .       .                                                                                                                     i 2         2      2 op(Fj ) = F y    r g + r-   .

(18) ' Because there can be random and systcmatic errors associated with sampling as well as analysis, the term 2r , equation (18) is frequently expanded to the form: 2 2 o{p),p p rh+r2+r . (19) 2 Similar equations could be written for of and oq , Equation (19) is valid for the case where the random error in F, is obtained from one weighing, one sampling and one analysis. There are some cases where mnitiple weighings, samplings, or analyses are performed. In such cases, equation (19) is not a valid representation of2e . If Ry weighings R 3samples, and RA enalyses are perforraed, it can be shown that: 2 2.2 2 o,(pj ) , p 2 + - (20) I Equation (20) is valid for evaluating the variance associated with the random error component. Since multiple weighings, samplings or analyses . have no effect on the systematic error component, equation (19) remains valid for the propagation of both long- and short-term systeraatic errors. tA , s

                                           .                                                     ;~...
                                                                                                                      , p3-.               -                                  -
                    . .          ..                                 ;;                        . . ;::                  ~ M*a-              %Tr?:.
      ,.  ~ . . :. = ~.           * -

a. .3 ' ,.:; - .(g . Q L Q -f '((

 ,e
      ,.n._.--                           _ - .        . -

h

l

t I

I,f l it Equations (19) and (20) are used for the propagation of systematic ] and random errors associated with individual flow or inventory terms.  ; The material balance requires a surrenation of flow and inventory terms.  ;.,I The propagation of systematic and random error variances throuch these summations are described in the following paragraphs. The behavior . of the flow terms will be derived first. This will be followed by an i evaluation of the inventory terms.

                                                                                                         '[

The analysis of the flow terms considers a neneralized flow term denoted by Yg . 5 Let Y .: g = u + cg + Og . (21 ) 3 Where u is the true amount being measured, cg is the random error made in the measurement of u, and Of is the systematic error made in the measurement of : . During the mg accounting period, assume n batches associated with the flow component "Y." In addition, durino the "m" accounting period assume the systematic error term 0 is a constant for all n. No distinction between long-and short-term systematic errors will be made at this time. If the sum of all Y{ is a measurement which is part - of the material balance for the accounting period, then: n n n n l bY 0. l 1 g=E i u + ib c4 + [i ( 22) 5 *

   .           Since u and 0 are the same for all n,                                                           .

l g.

                                                                                                                .I w

t g , ,em. "

,e

                                                                                                             ~
   . . - . - ~ . . .

i hY,=nu+no+ cg. (23) i . The variance of a constant times a random variable is equal to the variance of the variable times the constant squared. For statistically independent variables the variance of a sum is equal to the sum of the variances. . The variance cf a random variable plus a constant is eaual to the variance t of the random variable. Thus: o( (Y ))g = no +no. (24) i 2 The variance of all c, has been' expressed as o and the variance of 0 has been denoted by2c . From this equation it can be seen that if only one batch of material is associated with the accountino seriod, then there is no distinction between random and systematic error. On the other hand, as the number of batches included in an accounting period becomes large, since og and o usual y differ by less than a factor of ten, then the O contribution of the systematic error term far exceeds the contribution of the random error tem. Historically systematic errors don't persist indefinately. Let 0 in and Oq , representing short-equation (23) be broken into two comper.ents. Op term and long-tenn systematic errors in measurement respectively. Let o persist indefinitely but let the value o for p the mth, accountinq period g be taken from a distribution with a mean of zero and a standard deviation of o . Then by analogy to the previous development, the ' total flow variance for m periods, each with n batches becomes: i ,,,2 o [J = (ns.) o O2 , ,,2,2 , . (25) . 1 Y),j/ q p When the o's are expressed as relative errors, the equation for the feed , i term becomes: 1 1

      **-+=+-*...m              #
                                                                                .                " ~ ~*:~; .

y = y ' .* . - -- t uw . v.s i.s. A gg gm 1:s, .

 , e I

m. . _ _ .--. _ - -.--

_ _ _ _ _ - u n.w , m , _ i 1 n ) l; l o lf[m [F l (j i d'I/ PF r 2) j (mnf)2 [2 (F

• T
• E),

                                           "l 9 (26)        .

q f. The tem E is the average amount of accountable material in a feed batch.  ! Thus, the produc' mn? is the total quantity of material feed into the material balance area during i accounting periods. d i

   .                 The numerator on the lef t side of equation (26) is one of the flow                     "

terms in equation ( 8 ) for o (CUMMUF). Similar expressions can be derived for each feed and removal stream entering into the variance calculation. " From equation (26) it can be seen that the effect of the short-term I systematic error component on CUMMUF can be treated as if m values for the short-tem systematic error were randomly chosen from a oistribution with . 2 . a mean of zero and a relative wriance equal to pp . Thus, for one r, accounting period o(CUMMUF) is unaf fected by the fraction of the error that is short-tem. It follows that a(MUF) is unaffected also. However o(CUf91UF) is influenced if m becomes larce. The product en can be con-sidered as a time tem foe a facility operating at a constant throughput. Over a fixed time interval, since mn is constant, the tem q +r 2/mn is con-s tant. However, the tem pfm becones smaller as m increases. Thus, over ' a fixed time period o(CLM!UF)/mm is minimized by maximizing m. The tradeoff's between m and n will be demonstrated in Appendix B i for a 6 -r on th and 3 -month inventory in a reprocessing plant. Over f 1 i an interval of 2 years, the same total number of batches of material ' flow through the facility. In each case, however, recalibrations are g I assumed to occur whenever an inventory measurement is made. Thus, if an inventory measurement is made every 6 r.;onths there are only four ' t calibrations in 2 years, whereas for the 3-m nth case there are ' eight recalibrations. It will also be shown in Appendix B 2 that p is s. k

                                     *                                                         ,a.. .- .,

er l 2 large relative to q +r /mn. Thus, there is a strong incentive for maxi-mizing the value of m in order to minimize the value of o(CUMMUF). - The variance tem associated with the inventory terms in the MUF equation will initially parallel the development used for developina the variance equations associated with the flow tems. First, the . variance tem for the inventory term in equations (7) and (8) will be broken into random and systematic error comnonents. The distinction between long- and short-term systematic errors need not be made because MUF and CUMMUF contain only estimates of the beginning and ending inventory 2 levels. In both the a (MUF) and o (CUMMUF) equations, the inventory variance tem will be described by: o2(BI-EI) = o2p{g8 ,gE + I 'I ).

 .                                                           B E                                    (27)

The tems in parentheses in equation (27) should be taken symbolically and not algebraically. Thus, c2p{gB'I E) is taken as the random error variance of the inventory tem in the MUF equation. The subscript p denotes the systematic error variance of the inventory component. Equation (27) is used to describe the inventory contribution to the variance of MUF and CUMMUF. It may be argued that the "q" subscript, denoting a long-term systematic error *.ariance, should be used in the 2 equation for o (CUMMUF), such a distinction will not be used in this evaluation. There may be many categories of material on inventory, if there are k categories, all are assumed to be independent, thus k 0 2(gg _gp ) , g ,2p (78 -IE ). (28) i i i - l e

                                                         ~                                   _     q.                .

3 3'p g.

                                                                                                 ;"ptt;' .
                                                                                                                }
                                                                                                                ~..s q.y
         - ") .                         y~             ~

5.T

                                                                           , ~. .:3. . .:      {

1

        -v...-            . . _ . . . . -

3 ,- A similar expression can be written for the systematic error component. j The behavior of the measurement error variances associated with N inventory measurement cannot be surnarized in a form which is as simple [j 4 as the flow variances. Random errors propagate in a straightforward manner

                                                                                                                                      .i .

t,ut systematic errors must be handled on a case by case basis. Assume  : 4 there are k categories of material on inventory at the beginning and end ' of the accounting period. The inventory level in one of the k categories 3 1 3 may be best described by either an inventory level or an item count  ; times the average amount of accountable material in each item. These .I two inventory categories snow a different error structure as developed '$ in the following paragraphs. 3 d For the case where the inventory level is best defined as a total .- t quantity of accountable material, let I and I E represent the anount O B 9 9 of accountable material in the ith inventory category at the beginning and end of the accounting period respectively, fhen the random error variance i term can bs expressed in the furm: a g9 2 r i 3t),g932)r2 i i 1 i (29) a The systematic error term will be assumed to have tne following form: 2 { o (IBg 'IEi I

• U .~IE.)

B i 1 P i , (33{)2 p2 i (30) Eluation (30) assumes the relative systematic error is proportional  : to the change in inventory it vel rather than to the absolute inventory 1 - , l 1evel in the storage vessel. '-

                                                                                                                                   , e.

For the (ase where the inventory is present as countable items in an inventory category, let T gbe the average amount of material in the i,th category and C and C be the number of items on beginning and ending B E 9 g inver. tory in the 1,t_h, h category. Then the followir.g expression are used for the rar. dom and systematic error variances associated with inventory. *

                                                                                                                                      .a $

s

k. .-

                                                               ~ ~ .    '
                                                                           ,'      ~.f (
                                                                                               . . ,, / ** Y'e . . )

y^ .,a ' ,,y';h'?

                                                                                                         -                  .g: ,

 ,r 0 2{gg  ,gE)"UB +CE) p g       q      g        $

I3II o 2 (39 _3E)*( g g g (32) - In this case, the systematic errors associated with material present on

                                                                                                                                 ~

i ending inventory are assumed to not cancel with the systematic errors of i I measurement made at the beginning of the period. This is a conservative . assumption since some cancellation is likely to occur, particularly for ' short accounting intervals. For each of the k categories of material on inventory, either equations (29) and (30) or equatien2 (31) and (3?) ar: esed to proota;te inyt:nt.1ry errors. Equations (30) and (32) are very different. The evaluations per-formed in Appendixs I; and C assumes that equation (32) is the oroper fann whenever the inventory can be represented as batches of material which turn over during an accounting period. Green pellets which have not been - fired fall into this catenory. During an accountinq period, it is highly unlikely that green pellets present at the beginning of the accounting period will not be processed and replaced by new batches of unfired pellets during the period. Equation (30) is assumed to be the proper form whenever the inventory is present as a large batch of material. A storage tank con-taining several hundred kilograms of plutonium is placed in this category. If - a small amount of material is withdrawn or if the tank is emptied and filled , back up to approximately the same level, then it is believed that the systematic error component is proportional to the difference rather tnan the absolute . , s inventory of material present. .

                                                                                                                                    . l k'                                                                                                                                     l I

w, ,.M. . c. p. .g. g + . .

   .h[h hk?h$VtN.dWNk5NWh'&'k&_h_WOl.R;.hhj. hdh$'3
        = . . ..q_. ._ y
                    .         3-              . ;;.;z     . .g ; .          g.

y. sg gy .

    ,e w w . - . - ..         _

t I t j Equations (29) and (30) are applied to containers which are under item control and undergo processing during the accounting period. Items h that are on inventory and do no; undergo processing are included in the a MUF equation but not in the estimate of the a(MUF). Any errors made in 1

                                                                                                                   )

estfruting the amount of accountable material in these items cancel out j since they are present in both I B and I i E* l 2 2 The estimate of o (MUF) and a (CUMMUF) requir. .1 sumation of the k b categories of material on inventory. For each category, a decision is e made as to whether equations (29) and (30) or (31) ard (32) are applicable. For each class the variance term is calculated and then the variarice of all categories is summed to get the random and short- and long-tern systematic error inventory variances. These inventory variances are then sumed to ( 9et the inventory variance tem in equation (27). A.4 SU'4f1ARY OF NUCLEAR MATERIAL ACCOUflTING EQUATIONS The follt.fing equations are used to evaluate the sensitivity of nuclear material accounting netnods. -

1) MUF = 81 + El + F - R t 2) LEMUF = 2 o(MUF) c2{ggg)

         /                     3) a(MJF)     =

2

4) c (!!UF) = c2 (BI-EI) + c2 , ,2 ' ,

For n feed batches during the accounting period

5) c 2,72 nr 2 + n2 (p2,q)2 1 -

For f. removal batches during the accounting period

5) c 2=R2 tr2 , g2 (p2q) 2 i

c.-

      .t', $

W

 ! /i ,i

\

  !.b5.2::2W na:t:'
                                 ':%:b r         '

i4 - ' 'M ' W ~

                                                      .+. w:= wwsm                             -u      . .=     ~

,e

                                                                                          -   a
           ,.                 ...-.........,..,.,........_m._..._.__

34 If there are more than one feed or one removal stream then equations (5) and (6) must be developed for each stream. - The variance of the beginning and ending inventory term is broken down into k categories of material for each cateoory, a decision is made as to . whether or not it is describable as ah absolute inventnry level or as a series of discrete items. If the first h of the k categories is defined - by a total inventory level and the rest by an item count, then 2 2

7) o (BI-EI) = o,(g ,g E ) + " IB~I E )'

and h k 2

8) c (gB_g E ) " b IIB *I IT +bI B +CE I i=1 i i i i=n+1 i i i h k

9) 0 2(Ip B 'IF)

• AI I

                                                                                +{1         (Cf+C2   i )Ti p22 1-1          1        i=n+                        i, The equations for CUMMUF parallel equations (2) through (9). Equations                                                              ,

(5) and (6) must now cor. sider the effect of m at. counting periods. They become: 2 2 + n2 ,p2 + n2,29 SR) o (CUMMUF)s T mnr 2 p 6R) o2(CUMPUF)=li2 mnr 2 R + " "P

                                                                                              +""4

• All other equations remain the same. '

Equations (1) through (9) are used in Appendix B to quantify the effect of possible improvements in material accounting. Th'.t results , of these analyses have been summarized in Section 3 of this report. 0 _E k- +y-

                                                         ~ ~ '

N[_ ~ ' ' ~ *

                                                                                                                                 ;Q           pff.
                                                          .            :.-                     ,,              , .:= .           c=em .
  -.2 '
        ~~    . . .-;' *;            ^

TQ2; . d} ;'- ., .[  :-  ; .  ;.

me 1 l 85-REFERENCES

1) A. M. Mood Introduction to the Theory of Statistics, McGraw Hill Book Company, Inc., New York, 1950, page 247.

• 2) W. H. Beyer, ed., Handbook of__ Tables for probability and Statistics, The Chemical Rubber Co., Cleveland, Ohio,1966, paces 117-124.

3) K. B. Stewart " Statistical Techniques for Enhancino the Role of Material Balance Accounting in Safeguards," BNWL-1386, Battelle-Northwest, Richland, Washington, May 1970.

4) William Volk, Applied Statistics for Engineers, McGraw-llill Book Company, Inc., New York, New York 1958, p. 141-143.

t e e

                                 ,                                       ,   . a. .     -*~

me l l l i

               , =-                                                 .

1 l APPENDIX B

                                                                                                                                                      ~

SAFEGUARDS MATERIAL ACCOUNTING CAPABILITIES OF FUTURE PLUTONIUM PROCESSING FACILITIES This appendix will describe the detailed measurement uncertainty cal-culations performed to evaluate the sensitivity of future plutonium pro-cessing plants to detect material losses. Two processing plants will be

                                                                                                                                                    ~

evaluated; a 1500 MT/yr LWR reprocessing operation and a 200 MT/yr mixed-oxide fuel fabricatioa facility. Because of the vast differences in material accounting characteristics within a reprocessing plant, the model will individually consider the accounting characteristics of separations area, the plutonium nitrate storage area, and the plutonium nitrate-to-oxide conversion area. This evaluation will treat each area as individual facilities. A discussion will be limited to the plutonium material account-- ing capabilities of these four facilities. It should be recognized that the reprocessor must also account for the uranium processed through the plant. This appendix will completely develop the material accounting capa-bility of each of the four plutonium processing facilities before intro-ducing the next one. Each facility will be introduced by a brief descrip-tion of its operating characteristics. This will be followed by a discussion of the accuracy of accounting measurements performed for the facility. A final section will describe the material accounting capa-bility of the facility. In general, the last section will develop the . quantitative relationship which exists between the operating state of the facility and the timeliness and sensitivity of the accounting records. ,' The mathematical models described in Appendix A will be used to obtain . the material accounting capability of each facility.

                                                                                                   .               g                   4
                                                                                                                                  .g

=._  :.: '

r +.  ;.x n ,ve - ."mwg* mw;~ L } ^ "" ;r . , .'

 . 2 iJ^                                        I "' . ' J

{ g* ; =_.g 2 g"llllSl$.*.;;r.: . - . 7 .} . .

    ,r i
    ?
         ~ .. ,

B.1 MATERI AL ACCOUNTING CfsPABILITY OF TIE SEPARAT!01:S FACILITY DT"X 1500 III/YR LWP. REPROCE551HG PLANT As described in the previous paragraphs, the reprocessing plant has been divided into three separate facilities, a separations facility, a plutonium nitrate storage facility, and a plutonium nitrate to oxide conversion facility. This section will discuss the material accounting characteristics of the separations facility. The other facili-ties will be the major subject of subsequent sections. The separations facility operation is remote, performed behind many feet of c.s. rete. This analysis will be modeled after the 1500 MT/yr Barnwell Plant being constructed by Allied General Nuclear Services (AGNS). The facility description will be taken from the Safety Analysis Report (I ) prepared by AGNS personnel. It is thought to be representative of future separations facilities. Although the separations operation properly begins when the spent fuel is unloaded from the large shipping casks, from the standpoint plutonium accountability,the accountability tank which receives feed from the dissolver is the start of the separations process. The rationale for starting the plutonium accountability with the dissriver solution is rather straightforward. All plutonium in this stream is potentially usable mate-rial and will be feed into subsequent processes. Thus,it is properly the start of plutonium safeguards concerns. In addition, up until the fuel is

 ,              dissolved, its plutonium content is only known through calculations. Thus one works backwards from the dissolver solution to the plutonium content l ,              in the fuel assemblies used to make up the dissolver solution.

By starting with dissolver solution, one potential plutonium stream t l 1s not considered in the plutonium accountability calculation. This is ~ l i

                                     ,                      7.~,       . . - - +      - , ? e .- v n - n 6~ + - ~*+m='m**~~

y. 7 , - - = . , b*- .2-* [f

                                                                                                                     -    i
             .                      i- E f N 9 n-1 7 .7.'] T U 2E r~; ~ 3

• G ,'T. 9 E W z "OSFI

,r

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

the plutonium associated with the clad hulls. Since the hulls are pre-sently considered to be a waste stream and undergo no'further processing to recover the traces of uranium and plutonium present, the amount of plutonium, although of safeguards concern, has no impact on the amount of . plutonium which must be accounted for after it is processed through the separations facility. The input accountability tank represents the first point in the opera-tion where the uranium to plutonium ratio can be accurately determined. The uranium content is known quite accurately from knowledge of the initial uranium content in the fresh fuel and the subsequent fuel exposure level attained at reactor discharge. The dissolver solution from one of the three dissolvers is batched into the accountability tank and then jetted into the feed tank for the first column in the separations procedure. Once the solution is transferred from the account tank it becomes part of a continuous process and batch identity is effectively lost. The Furex separation process is shown schematically in Figure B.l. All chemical processing activities beginning with the transfer of material from the dissolver accountability measurement tank and ending with the transfer of product material to the storage area are included in one material balance area. . There seems to be little incentive to divide the area into more than one material balance area. First of all, except for the final concentration . In addition, all flows are and storage steps all operations are remote. cor.tinuous or semicontinuous with recycle and backcycle being used exten-sively to obtain the desired product purity. Thus multiple material balance areas bounded by reliable material measurements would be difficult to

                                                                                                                    = . .
                                                                                                     , 'NN '    4 ,.
                                                    ' ~ '         -
                                                                ' '}~            c'*Q-
   ' ci ~

597 ,: y ;p, y.- jQ- . , . . -

                                                                                                    ,W4ref
                                                                                                              .;Q
                                                                                                                          ~.

2 T I I

                                                                                       \g 2nd U d       Extraction       H U

Silica Gal (j Cycle Nitrate l Purification Prodact Of f-Ga s a " i ': To solvent Spent fuel Fuel Chop treatment receiving (1) 1st Solvent Extraction - s Dissolution 4

                                                   & storage            '                   Cycle - U/Pu Partioning                       if
                                                                                                                          ~
           "                                                                                                                             To Low Level
                ,                                                                                               /N (2)                                               Waste Concentrator v

Hulls to Interim II U' 2nd Pu 3rd Pu (1) Pu q tion q Ex r E ton q Nitrate Storag.- (solvent) 3, s, Treated solvent for reuse V + Solvent Misc. (1)* Primary Measurement Points Trea tn:en t Plant . for Plant Material Balanca us;ns i P (2) Leached Fuel Hulls are I sampled on a statistical High Lt.wel_ basis. A nominal loss of Waste i Low Level qH O 2 for reuse 0.1% of the feed U & Pu is expected. N Concentrator Waste ((concentrate [sron:entrator

                                                                                                                                                    - -)HNO3 for reuse r

(1) Interim On-site

            .                                                                                          y         Liquid Storage
            ,   :                                                      _ FIGURE 0.1 Conceptual Flowsheet Purex Processing Plant i

7y I, O y .. .__a. :n *

                        ~ _ - -

_ = --~._= _-. -. 2 I 1 B.1.1 Description of Important Material Accounting Characteristics of a separations Facility . The material accoun'ing characteristics can be strongly influenced by whether the plant is maintained by remote or contact maintenance. Of rectly maintained equipment must be designed to be thorcughly

• decontaminated. This aids material accountability. At the sarre time, the requirement for decontamination lengthens the downtime. A minimum outage would probably be 10 - 14 days. Maintenance outages would probably last at least 30 days. This means that although the facility can be almost flushed clean, a requirement for frequent material accountinos would probably result in lost production.

This analysis will follow the Barnwell design philosophy. All but the head end operations are assumed to be maintained using contact mainte-nance. Formal material accounts will be assumed to require thorough flushing to remove the major portion of the plutonium from the remote process equipmer.t. Present regulations for separations facilities require material account-ings be taken at least once every 6 months. Table B.1 shows a typic mate-t rial balance for a 6 month period. The plutonium content in the feed stream l was assumed to be 10 kg/ tonne of initial uranium in the fresh fuel. This I is above the expected concentration expected in discharged uranium fuel but below the concentration expected if recycled plutonium fuel assemblie.s are being processed. The plutonium losses in the high-level waste stream represent 0.9% of the feed plutonium. Lower losses are expected in practice, a high value insures that the estimated measurement errors for the waste ' stream will be conservative. Although the data presented in Table B.1 are for a specific accounting ~ period, different accountability intervals are easily obtained from Table B.1

                                                  - . ,      .~           r 4

ff " hi 3 $. . [._W.

               .__ ,_                              ~.         a.:         == :       WQ            3   _}        - ~W.

Z3f-j~ rj

                                     .?         2'E~ ;

u 37Q'{ - 'W M ~ . M%,-- TMT

                                                                                                                                                            ~

3 [ . I ' IABLE B.1 f Material Flow for a Single Campaign at a Purex Processing Plant (Basis: 750 Hetric Tons Fuel Input, 6-Month Operation) f 4 _ Material Balance Component Batch Size kg. Total Material (a) No. of Batches in Campaign, kg. Total Input Pu U 20.182 375 - 6743.25 1998 375 749,250.00 Total Product (b) Pu U 25.0 300 6682.5 7425 100 742,500 Waste " Pu $. O.273 250 ' U 60.75 27.0 250 6750.00 , In-Process InventoryI " Pu U 0.5 10 5 5.0 10 50 a) For this example, it is assumed that the concentration of Pu is 9kg/ ton of U. In actual operation, this value can be expected to range from 5 to 10kg/ ton of U. A nominal 0.1% loss with the teached hulls is assumed. b) Pu at 200 g/t; U at 1.5 molar, c) Inventory assumed for a " clean" plant. Small amounts of U & Pu in process are trar.sferred to tanks where measurements can be made. The inventory in a " clean"

                                                                                                                                                                            ~

plant is espected to be'nearly constant at each 6 month inventory period. n 1 1

i

                                                                                                       ' ~~

w w -~~ , . by applying a fixed ratio to the column specifying the numbered batches pro- . cessed. This procedure will be followed to evaluate the relationship between sensitivity and timeliness of material accountings for a separa-tions facility. This information is developed in section 8.1.3 of this appendix. This section is proceded by an evaluation of the accuracy of . present measurement techniques to detennine the quantity of plutonium in each stream entering or leaving the separaticns facility plutoniua mate-rial balance area. B.1.2 Measurement Uncertainties for the Separations Facility Plutonium Accountability Heasurements Measurement uncertainties for two types of account-ability measurements will be developed in this section. First, the measurement uncertainties associated with what are considered to be the best formal accounting measurements will be described. Then the estimated accuracy of running inventory measurements will be described. There are many years of experience with formel material accounting methods for plutonium accountability in a Purex-type separations facility. Perfo. -.nce data from facilities in this country and in Europe can be used. As a result, the present capability of measurement systems and techniques has been extensively documented.I2-U Table B.2 represents a com-These numbers should be posite sumary of estimated measurement errors. considered to be representativt of present perfomance. -They represent

    '               measurement errors associated with what are thought to be the most accurate measurement methods. Analysis of the plutonium content in the input account-ability tank is obtained using isotopic dilution techniques to get the uranium to plutonium ratio. The uranium content can be accurately deter-mined from knowledge of the initial uranium content and discharge exposure

~~ & .~; 9:- y  ; &Q p:g. .g  : .2 ,g; y.g g

   ;=+ - n=.           -

4_  ;.3

                                                                                    --   .-_          _ z ~:.    ,

- . . . l i

l' i -- TABLE B.2 r Stimates of Random and Systematic Errce for Separations Facility Plutonium Accountability Heasurements Material _ Balance Component Relative Percent' Standard Deviation #' of a Single Measurement g Plutonium Input Volum5~ Sampling Analytical - 4 e

                                                                                   + Random tQ'

• Short-Term. Systema tic

                                                                                                                                +

0.3 0.30 1.0 Long-Term Systematic 0.18 0.10 0.20 0.02 0.02 0.02

       . '[

Plutonisp Product y s ~ D' Random

   -ji /                                    -

Short-Term Syst'ematic 0.3 0.1 - 0.2 Long-Term Systematic 0.1 0.1 a-0.1 - / - Y 0.02 0.02 - ' ~ 0.02 Plutonium Waste _

       ""                                                                             Random Short-Term Systematic                           2.0     .-

6.0 20.0 s Long-Term Systematic 3.0 6.0 1.0 10.0 1.0 1.0 - Plutonium Holdup

        ;                                                                           Random / Vessel                                                                                  .  /

5.0 5.0 , S.0 - b s J l y t

             .                                                                                                                                                                                                                  a t
                                                                                          ..P"'

l l

            ~.c   .        --        . . . . . .
                                                                                                                                                                                                     ~

of each fuel assembly in the dissolver batch. The plutonium content in the 8 product is determined by coulometry and in the waste by TTA extraction-alpha , counting. These analytical methods when used in conjunction with good volume calibrations and sampling techniques provide the best estimate of the capability of future separations facility measurement methods. It is fully recognized that some facilities will exceed the measurement accuracies  ? reported in Table B.2; others will not quite meet them. A factor of two deviation fran the accuracies is possible but larger deviations are thought to be unlikely. The systematic error term has been divided into a short-term and long-term error component. At present, no studies have been perfomed to quantify how much of the systematic error term will be reduced by the measurement l control program. The 0.011 long-term systematic error represents that fraction of the systematic error that may persist over many accounting periods. As shown in Appendix A, the distinction between long-term and short-term systematic errors does not affect c(MUF) but it does become important in evaluating o(CUMMUF). Table B.2 sunnarized the accuracy of preseret formal material accounting

     -                  measurement methods. Because shutdowns for physical inventory measurements require 10 - 14 days at a ir.inimum, running physical inventories must be considered to be a serious alternative.

Based on the Barnwell design, which is thought to be representative of future designs, most of the inventory is present in feed or accu.mlator tanks. Column 2 of Table B.3 presents a compilation of the estimated average inventory of plutonium present in the flowing streams during normal operations. The estimated accuracy of measuring these inventories is shown . in subsequent columns in the table. The accuracy of inventories in the , separations columns was taken to be 105. The actual inventory level and measurement accuracy in an operating niant would be obtained by experience.

                                                                "                           ~      .~.,.           ..
                                                     ,M.*.   ' *y ? *'

f _-, f, _.

                                                                              '~t* ~3 a   #ln     .         ..
                                 . 5 -.       &- - - -

d~&:.  ;&f

                                                                                          }   .-        . _ -

f- - -

7

4. '

TABLE B.3 Running Inventory Measurement Uncertainties for the Separations facility of a 1500 MT/ Year LWR Reprocessinn Plant Plutonium Component Es tinated Measurement Accuracy Process No. of Inventory  !. l Component Components Fg Volume Sampiinn Aria ly t i ca l Fuel Dir. solution Accountability tank 1 23.13 Ran. dom o.3 0.3 0.1 Accountability tanF 1 23.13 Sys te- 0.18 0.1 0.20 matic Centrifuge 1 1.5 10(,) HA Feed Tank 1 21.5 1 3 0.17 Flush Accum. Tank 1 6 1 3 0.17 U-Pu Co-decontamination Cycle HA Column 1 1 10 HS Column 1 1.2 10 IB 1 0.5 10 IBX Column 1 0 10

                                                                                                     ~

Secondary Recovery 3 <<0.! 1 5 0.17 Plutonium Purification 1 BP Feed Tank 1 1.7 1 5 0.17 2A Column 1 0.5 1 10 3A Column 1 1.3 10 33 Column 1 1.3 10 2 PS Column 1 20 10 2 P Column 1 11 0 10 Plutonium Catch Tank 1 1 5 0.17 Plutonium Rework Tank 1 7(b) 0 25 Plutonium Collection & Storace Pu Sample Tank 1 21 0.32 0.14 0.22 Pu femp. Storage Tanks 3 42 Randum 0.3 0.10 0.20 , Sys te- 0.1 0.10 0.10 t matic Pu Temp. Storage Tanks 3 -42 Fh nd om0. 3 0.10 0.20 Syste- 0.1 0.10 0.10

                 '                                                          matic Pu Measurement tanks           1              11              0.32        0.14        0.22 a) If only one error term presented. that value indicates the total measurement error used.

b) Rachig Ring Filled Tank (Capacity 200 kg of Pu) assumed not used *

      .                   during accounting period.
                                                                                                                     ~

9 e g , g p. *@T @ "M' #

            ~

The values are thought to be typical. The biggest error associated with the

      # ~ ~'~    .

running inventory measurentent is the inability to guarantee that a sample taken fran an accumulator tank is representative of the plutonium concentra-tion in the tank. Sampling errors below 51 may be impossible to realize. in practice. - The inventory levels shown in Table B.3 are based on the plutonium concentration in the flowing streams. In the high-acid Purex flowsheet. . plutonium deposition from solution cnto the walls of peccess equipment is not thought to be a major problem. This fact must be demonstrated in practice. 8.1.3 Capability of Material Dalance Accountino Systens for Plutonium in a 1500 MT/Yr Separatinns f acility Based on the results presented in the previous sections of

     \

this appendix, it is now pcssible to use the error propagation models described in Appendix A to obtain the capability of material balances performed over a separations facility. Three cases will be considered. First, the effect of the accounting

 !                             interval on the value of LEMUF will be evaluated. This will be followed by the results of the running inventory evaluation. The third case will
          -                    evaluate the possible effect of the measuren.ent control program on the long-term reduction in systematic errors.

The variation of LEMUF with the frequency of the formal material accounting interval is shown in Table B.4. This table clearly demon-strates the behavior of LEMUF with throughput as discussed in the , introduction to section 3. Although the value of LE:1UF increases as

                         ,      the throughput increases, the value of LEMUF, expre;;ad as a percentage                          ,

of the feed, decreases. The former term is a criterion for evaluating , j

                '              whether a single large loss has occurred whereas the latter is a criterion                            l for evaluating whether small frequent losses are occurring.

i Although the table shows formal accounting intervals as short as AdE" it

                                                                        ~

r =t R1 &

          =.:      . h. -E.
                                     =.=.-              . :si. x. . ..
                                                                       , -6 =
                                                                                    .5_. :       Ah j'.=%

WQ

i9 po"" i, a 6

                                                     =                                                                                   =*-                                   ==

e eO. enN.8* o  %. O. %. p og. O e eftU e. . e ** N. O en O 8.s

                                                                                                                                                           . o.        e. t em.

en e% =m c O es e e as % sen 8 e O em os as O em. em e Sh e ee eO em e-upt N ene et e em

                                                                                                                                                        %%                                                                 9 9                                             *

e. ce == et T e ce se e no e O. e. e. e. e. o es g .e 3 e s e m os d ee m s e.est. s't. N. ets. e.

                                          &                                     *******                       D       w em O c am e           p. ee e         o et.
                                                                                                                                                                           ,0 8m.
                                          $       en    up                      80            ow            4't P

eft ee st et sie # O e

• om - M et'h  %

e +> ., s

                                          .e g          e=       4     *.

b e* e, s es e O sq3 m e e OeeO est is o  :. * .O. e. e.n s'.t es.ar.e g o a = e c-aee O- ee w O. w. e. e's. em.N.F. e e. e on. elf M w 6 e eat == e.O ew D ee ms e es est e em - e.s se a= e- e. e e.t e = to C = 6 8 eme m ' G- . . gL - es M 6 .a. e.== es e et e a e =O.- en e w sit e. gg m s u e e c. %. em. e. em. e. 8'.t ee. er.i. ** e. em. se.e. s.os. e o meae OOwe me** o O *e's a 9 ne$ a

                                                       =

9

                                                                           -= **                          -

a -- e ene e seetc - M e w w CG s a OU "en . 8' e er= 0 e a eOe wom a% .sa e we. Oee wn 9 O- 3 +=.O.

                                                                                          .                    w. e. n .          m. . v. e. oe. s . em.            O. c. ee.

3@ *s ee ** e O O w oft - O - se

m. me ==e= est e ** eveft *OO

                              *M                                                                                                                     one me              ==

v e'"3 3m CK = l=J 3 I w = e- e v e. ,

         /

g een a s

w. O. e. ee

                                                                                                                      .s.

e. n O Cee w e e.

O. w. e. e. .e. o. o. o M b e. O-Me art O - e= .6OOO -

                                                                       **                               *e                        es                 e e4 e % eP O C #45
                          **~ G m >=

M *m ms b , a s es e - m es en - , et e e m .e; o. em. em. e.eft O. sa. O. en e*. wt e.st O. O. 9 U a * .- e e . N.e. OOOO

                                                                                                        ~Oco ev e o         ee = =

CO e-

                                                                    "                                                            ~

20 nn un C L. O 3 we e sqe g -a ..m= g e em =

                                                                                                                                        .et             O at                                              - ..                         , e as                   a            m O          e                                          e e.

e9m e e o a e se % e4 w w as v. sur. eee e .= 3w at 6 O S h===T* O-= e .a . 9e es

                                                                                                                                                   .em age te b

edb, 8.e w e .b O se e y e= * > em. 6 e .en.n m .e e.mJ e .e m a ema oa-a gg*, -6me

• s6 . -66 e e. se e

                                                           -en m. se                       e                oe                     ee e en                                           ese                     =se              ee                     e Mi                                 e e b aus t ae =

bw 5 e" em b to e bw eee 9 se e es kw

                                                                                                       =o e b ese a.          .en e .b ese e          www                    =

em ee b .b. Jene 6se . w 6 go ses e  % Pe es b w b pg es b - E g44 O

                                                            + e ene es 9                               game me se                                                4 O es                            eo        e se                  en e ene me e           w--                eO e>
                                                              == o e en Se                            Ee             es.e ggea                b, be b

e. e e == e G = es e es == a * = Oe-Ot- *** *'* e e E em te s* O w as se e O en e em 2 E p as &

       $                                                   w S e Dn O                              *e O g em e            se e g hO W et as        eso e= em g4g 3 sh =
                                                           & E & treP=                             @ M M afth.             em 5 & eft ese         es****

E 3 *.B "EE O b g e b o e se O = 3 6. em De eft .t.e m ameene e e

                                                                                                                                                                                                                               -h e  s                                                                                                        g  .. ,
                                                                                                                                                                                                                   ,*"*'-N   ,

e e e 4

           , , ,                 _ . _           - - ~ - , ; .- ,-
                                                                                                      .,,n,",.T.___                                 ,i-,_                         --= "-                  n%'-       N~%..
                                                                 ,                                                          . _                                                              , _ . . --         + -

weekly, running a plant for a week followed by a 2-week shutdown for a

 ,                               physical inventory measurement is not econortically feasible. They are shown to compare with the runni'ig inventory evaluation developed in Tables B.5 and B.6.                                                                                  .

Table B.5 presents the estimated accuracy for a single running inven-tory in the separations facility. The accuracy of this running inventory . can be used in two calculations. Assume the accounting period is bounded on one'end by inventory cbtained using formal inventory techniquas, then the uncertainty associated with that invertory measuremant is very strell; half the value used for the invent (,ry term in Table B.4. The other inventory uncertainty value is associated with a running inventory. The second rase assumes the accounting period is bounded on bcth sides by a running-inventory estimate. Table B.6 compares these two cases with the best formal balance estimate. Based on this comparison, it can be seen that running-inventory material accounting techniques may provide the most acceptable assurance that all the material feed into the system can be accounted for at intermediinte times during a formal material accounting period. The running inventory is not thought to be a substitute for a formal accounting material balance. When the plant is shut down and relatively ' clean, it will be convenient to check the calibrations of the various measurement systems. This is required as part of current regulations.(9} The major effect of this measurement quality control program is to randomize some components of the systematic error. It will never be possible to completely randomize the systematic error between accounting periods. Tables B.7 and B.8 snow the effect of randomizing the short-term systenatic errors shown in Table 5.2. They are randomized by assuming new calibrations are used for each accounting period. Table B.7 shows the effect of recalibrations

                                                                             ~ .           -

N. c.--

                      .~                            -              .      .
                                                                                       .g         -y           3:            ,

IC gue::Q mm=

      .. n
         . + .
                   .a.

=

                                     1

_2~;

                                                             ..: L n; y -        g-~;      ..72
                                                                                        ;'_ '      ~Me         -g.g; .

TABLE B.5 Running Inventory Measurement Uncertainty Evaluation for the Separations Area in a 1500 MT/ Year LWR Reprocessing Plant Total Plutonium Estimated Measurement Component Measurement Error Process No. of Inventory Accuracy Va riance Component Components kg  % kg 2

   .                Fuel Dissolution Accountability           1                23.13                Random 1.1    0.063 tank Accountability           1                23.13                Syste- 0.26   0.004 tank                                                         matic Centrifuge               1                   1.5                      10.0   0.023 HA Feed Tank             1                21.5                         5.0   1.156 Flush Accum. Tank        1                   6                         5.0   0.090-1.386 U-Pu Co-decontamination d                        Cycle HA Column                1                   1                        10.0   0.015 HS Column                1                   1.2                      10.0   0.015 18 Colunn                1                   0.5                      10.0   0.003 IBX Column               1                   0                        10.0   0.00 0.028 Secondary Recovery           3            <<0.1                           10.0   0.000
                                                                                                                       ~

Plutonium Purification 1 BP Feed Tank 1 1.7 5.0 0.007 2A Column 1 .5 10 0.010 3A Column 1 1.3 10 0.017 33 Column 1 1.3 10 0.017 2PS Column 1 20 10 0.040 2P Column 1 11.0 10 1.210 Plutonium Catch 1 _ 7 5 0.122 Tank Plutonium Rework 1 0(a) -- 0.000 Tank 1.426 Plutonium Collection

                       & Storage
   .                    Pu Sample Tank           1                21                           0.41  0.008 Pu Temp. Storage         3                42                    Random 0.37  0.074 tanks Pu Temp. Storage         3                42                    Syste 0.17   0.047 tanks                                                         matic Pu Measurement           1                11                           0.41  0.002 tanks                                                                       0.131
                                                                                                                    ~~

TOTALS 238.6 2.921 a) In Barnwell -200 kg's of plutonium could be stored in this Rachig Ring Filled Column. No additions or withdrawals are assumed during the ri.nnir.g invantory period. b

• e, -- .ee. * *

   !? '                                  +

4 t m.

!        i' iR[i .f:

I i _ TABLE B.6 Comparison of LDtVF Sensitivity to Frequency and the Type of Physical Inventory

,);
'i Performed to Obtain the Material Balance for a 5 MT/ day Separations Facility LEMUF Formal Inventory           Running Inventory Period on One                on Both Sides Size of Accounting           of Accounting Accounting          Formal Accounting v!d           i Period                     Interval                     Interval 1-                           Period 4.91                         5.61 1 Week                   4.09 6.44                         6.99 2 Weeks                  5.85 19 ;-                                                                                                          10.77 1 Month                 10.07                       10.43 19.04                        19.24 2 Months                18.85
                                                                                                                               .L E

I

p. Ui t i ,. , ...

} bi 5 .. j ,)'t1_ ; O I N fh. ; 1

l 4

                                                                 -131-                                                          g
   ~

I t at 6-month intervals and Table 0.8 shows the effect of a 3-month recalibra- .

   .                tion interval. This pro; ram has the greatest potential for increasing the sensitivity of material accounts to detect small, frequent, undetected losses.         measured as the ratio of LEMUF to feed on the last line in the tables. Gains in sensitivity are significant when recalibrations occur every 3 months instead of every 6 months. This shows one, perhaps unexpected, advantage of material accounting intervals more frequent than the present 6 month reporting requirement for separations facilities.

B.2 MATERIAL ACCOUNTI"G f10DELS FOR THE PLUTOMlutt NITRATE STORAGE FA_CILIII The plutonium obtained from the separation facility can be stored as nitrate or sent to the oxide conversion facility. In this evalua-tion, all the plutonium nitrate is assumed to be sent through the scorage facility even though economic, safety and safeguards concerns may suggest otherwise. Two overriding reasons may result in tne extersive use of the facility. First of all, there is an advantage to the fabricator to obtain plutonium having the same isotopics in fairly large batch sizes. It greatly simplifies scrap recovery operations and, as a result,the fuel is much more homogeneous. The possibilit'y of removing americium from plutonium nitrate solutions is the second reason why plutonium might be stored as nitrate. Both the reactor operator and the fabricator like the plutonium assemblies 241 to be low in americium. The fabricator because of the Am dose to workers 241 Am is a nuclear poison. For both these and the reactor operator because reasons, fair amounts of plutonium nitrate might be stored. The Barnwell plutonium nitrate storage facility is thought to be representative of future facilities of this type. At Barnwell, I tonne of plutonium can be stored in six interconnected slab tanks. The present OM n

                                                 * ).g                     e  e           4       mOy   e=     w  +h   b w' , g

_ 6-. h

    .tn :

F r. : C i..?I 1: ,. P' TABLE B.7 Effect of the Measurement Control Program on the Lcng-Term Measurement Errors for Semi-Annual Inventory Periods Over the Separations Area of a 1500 MT/ Year LWR Reprocessing Plant Elapsed Time from Initiatica of the Control Program i' Q 6 Months 1 Year 2 Years 4 Years 8 Years

'J                                          Number of Inventory Periods            1           2                4             8             16                             .
                 -                                                                           Cumulative Measurement Error Variances                                        .

Feed - 20.182 kg's of Pu

    .F.                                        No of 84tches - 375 ff!                                        Randem Error Variance - kg 8 18.02             36.05           79.09       144.2            288.4 W l                                        Short Term Systematic          472.54         945.08       1050.18       4628.9           7560.8 Error Variance - kg 8 Long Term Systematic              6.87         27.49        109.97         439.9          1759.6 Error Variance - kg 8 Total Variance - Feed - kg 8 497.43         1008.62        1269.24       5265.0           9608.8
,lii;'1 % !

Product - 25 kg e No. of Batches - 300 Random Error Variance - kg 8 2.6 5.2 10.5 21 42 5 i Short Term Systematic 162.0 324.0 640.0 1296 2592 'y .

                .                                 Error Variance - kg 8                                                                                                    ,

Long Term Systematic Error 6.8 27.0 108.0 432 1728

    ..                                            Variance - ag 8 Total Variance - Product - 171.4              356.2          766.5        1749            4362
    !!! M j..

9. kg 8 Waste 0.273 kg No of Batches - 250 Random Error Variance - 0.82 1.6 3.3 6.6 13.1 kg*

Wid' y. Short Term Systematic 66.14 132.3 264.6 529.2 1058.3 Error Variance - kg 8

    'l.d.

Long Term Systematic 1.39 5.6 22.4 89.4 357.7 Error Variance - kq 2 Total Variance - Waste - 68.35 139.5 290.2 625.2 1429.1 kg 8

             ,1                                   Total Flow Variance          737.2        1468.3        3137.0         7647.2         15400 k

y. Inventory Variance 2.2 2.2 2.2 2.2 2.2 y;[ h> t-Total Variance 739.4 1670.5 3139.2 7649.4 15402 Sd(MUF) - kg 27.2 40.9 56.0 87.5 124.1 LEMUF - kg 54.4 81.7 112.0 174.9 248.?

    . . . .tja - :                                                                                                           0.29             n.20 M  gt                             (LEMUF x 100)/ Feed-1          0.72          0.54              0.37

' m, [lplh J t

1 t TABLE B.8 Effect of the Measurement Control Program on the Long-Term Heasurement

         <                                                           Errors for Quarterly Inventory Periods Over the Separations Area of a 1500 MT/ Year LWR Reprocessing Plant tispsed fire 'com Initiation of the Control Program 3 mo ntas               1 Months 12 Months  ~15 usnths in ac nths 21 montas 2 years L ucers              4 years      5 eears tw*ter of Inventory Periods            1 - 6 weaths 2           3           a        5          6           i           5          12            16          20 Cumulative weasurement trror Verlance feed . 20.18 kg 8atches to of Battnes/laventory Perlos . 187.5 Rensse Error variance      bg'      9.E ll    18.02       27.033    36.94      45.06      54.37      63.07       72.08      108.13         144.17      180.22
't.

Snort fera Systematte Error variance . eg' 118.14 236.28 354.42 472.56 590.70 108.84 826.98 945.12 1417.68 1890.24 2362.80 Long fera Systematic i Error Verlance . tgs 1.718 6.87 15.a6 27.49 47.95 61.85 84.18 109.95 ?47.39 439.at 6R7.20 Total vartence . Feed kg8 12s.67 261 17 "7IT3T 536 04 77C7T "T2T.F 914.23 1 27.15 *T77 T2T 2474.22 M s." Predect . 25 kg Satches No. of Satches/ Inventory Period .350 Ranson Error Vartance . 1.312 2.624 3.936 5.246 6.56 7.87 9.18 10.49 15.74 70.99 26.24 kga Short fare Systematic 40.50 81.00 121.50 162.00 202.50 243.00 283.50 324.0 486.0 648.0 810.00 a trror Verlance . tg'

             '          Long Term Syst* matte Error Verlance . kga 1.687     6.75       15.18     26.94      42.17      60.73      82.66      107.97      242.93         431.87      674.80      $
'n                      Total vartence . Prodett kg'       43.499     9 0. 3 '?  140.62    194.24     251.23     3tl.60     375.34      442.46       744.67      1100.86      1511.04 Weste      0.273 kg Sattnes No. of Batches 7lnventory                                                                                                                                              l Perted . 125                                                                                                                                                       +

Sandon Error vartance . ng' O.409 0.818 1.227 1.636 2.04 2.45 2.86 3.27 4.93 6.54 8.18 Short form Sestematic 16.54 33.08 49.62 66.16 87.70 99.24 115.78 132.32 198.48 264.64 330.h0 terar Verlance . tg' Long ferm Systeestic 0.34 1.16 3.06 5.a4 n.50 12.24 16.66 28.76 48.96 87.04 115.00 trror variance - ng 8 Total variance . vaste Ag' 17.29 35.25 53.91 73.24 93.24 113.93 135.30 157.35 252.34 358.22 C74.98 Vetal Verlance .' Flow . kg 189.66 386.79 598.44 803.57 1023.!3 1250.29 1484.87 1726.96 2770.21 3933.3 5216.24 Total Inventory Vartance .'6g' 2.16 . Total variance . tg a 191.82 288.95 593.60 805.73 1025.34 1252.45 1487.03 1729.12 2772.37 3935.5 5214.4

• Sd(MUF) . tg 13.85 19.72 24.36 28.39 32.02 35.39 38.56 41 58 52.65 62.73 72.74 i tin J F . te 27.70 39.44 4a.73 56.77 64.04 70.7A 77.12 83.17 105.31 125.47 144.48 I (ttstuF a IO0/ Feed) 1 0.732 0.521 0.429 0.375 0.338 0.311 0.298 0.47 0.232 0.207 0.19 I q.

4

  .s'.        .

i \

                                                                                                                                                                ........,..s.w--

e* .

l

                                                                     -104-storage facility contains four banks of six tanks each. Thus up to 4 tonnes                           '

of plutonium as nitrate solution could be on inventory at any time. Future expansion could double the capacity so that 8 tonnes of plutonium might be on inventory in the facility. The 8-tonne capacity represents the entire production of the separation facility for 7 months. As such. it does not . represent an excessively large surge capacity between the scparations and conversion facilities. , B.2.1 Description of Important Material Accounting Features of a Plutonium Nitrate Storage Facility Based on the 1 tonne plutonium capacity of each bank. several modes of operation are suggested. These are:

                                 . Banks of tanks in static condition where solution volume (or                                    l 1

weight) is essentially constant over the accounting period. l

                                 . Banks in which cnly interval mixing has takere place or from which only shipments (transfers out) were made during the accounting period.
                                 . Banks in which solutions of a different plutonium concentra-tion were received and mixed with previously stored material.

The timeliness and sensitivity of materials accounting checks will be somewhat different for each of the above conditions. The next section will i prmnt the basic measurement uncertainty data for each of the abava cases. This will be followed by a section quantifying the sensitivity of the h! accounting methods in the actual storage facility being considered. ll B.2.2 Measurement Uncertainty Estimates for the Plutoaf um i' Nitrate Storage Facility

..                                           The plutonium nitrate storage facility is considering only one

.)

. j                       material form. As such the relative errors for receipts, removals and stored                         -

t .' material are identical. It is assumed that the measurement techniques are the l .t , j same as those used for the plutonium nitrate product solution from the l

                                                                                       . . -         M?          . w.
                                                  ,y.       -e.~            ~* ,

w-  ;. R&h [, -j' I , .. _ h ^[$E .h __. -

                                                                           ;gh - . - --hh                -

5. _ ' - _- 3! -.

[

l

              . . . . .                                                                                                  1
                                                                         ._.                                             i
        ,                                                  -105-separations facility. These are the same measurement errors already pre-sented in Table B.2 and are used for all flow measurements in the storage '

facility. The inventory measurement error associated with material stored in the six connected tanks is obtained by assuming that one sample and one analysis is performed for each bank. However, the weight-factor (v'olume times specific gravity) reading for each of the six tanks is used to get the estimate of the quantity of plutonium solution in the tank. The effect of thisprocedureistoreducethevolumeerrorassociatedwithstoredmaterial by 4. Table 3.9 summarizes the randun and systematic errors used for the plutoniun nitrate storage facility accot.ntability measurements. B.2.3 Capability of Miterial Balance Accounting Systems for a Platoaiem Storace facility The material accounting capability of storage facility can be obtained by combining the uncertainties associated with the various operating modes which exist during an accounting period. The longer the accounting period the greater the complexity of operations and,as a result, the greater the total uncertainty of the measurements. Thus there is truly an incentive for relatively l short material accounting periods since the operations are relatively simple. ( The first set of cases will consider one to four banks static throughout the entire material accounting period. The second set of cases will consider the cases where 1000 1.o 4000 kg of plutonium is being mixed in the facility. The third set of cases will evaluate the effect of shipping and receiving l from 50 to 2500 kg of plutonium as nitrate. If less than 1000 ka is shipped

          .        or received, only two banks must be active. All four are assumed to be active
                                                                                                                     ~

for the case where 2500 kg is shipped and received during an accounting period. I i m . .e * ** '

l i Ybh'> *

                                                                                                                                                      +

, L97 ,4 $ I 1

• TABLE B.9 Estimates of Random and Systematic Error for the Accountability l

Measurements of the Plutonium Nitrate Storage Facility f pl! ' ') i i, Relative Percent Standard Deviation Material Balance Component of a Single Measurement

                     '                                                                                                                 Analytical
     ,;,                                                                                       Volume                Sampling                       f i
     -Q; ;

Pu(NO3 )4 Random 0.3 0.1 0.2 Systematic 0.1 0.1 0.1 i tii:. ;

,  ((jjj                               Pu(NO3 )4 Random                                                0.12                   0.1                 0.2    3 0.1                    0.1                 0.1    g Systematic 1
  -Q; .                                Pu(NO34         I Input to 0xide Cony.

I! Od, Random 0.3 0.1 0.2 0.1 0.1 0. Systematic l Material Heldup in Empty Tanks SO

              .,       !                   Random Error / Tank j ,,I N .P. (m: . !

4 9 2 di Uf ry' I di j' l

                                          .S                 4                                                                .
        $                6 4@*+5 e5D O v

                                                                                                                                      }

i

                                                        -197-Case 1: Material Balance Capability of Static Tanks                                                                .

If tanks are static, the sensitivity to detect loss is ' independent of the time interval between accounting periods. There are no measured flows. In this case. the volume (or weight) in each tank is checked at the beginning and end of the accounting period. No samples or analyses need be performed. Thus the variance of the resultant inventory measurement is only affected by the variance of the volume readings. For each tonne, six ! readings (one for each tank in the bank) are taken at the beginning of the accounting period and six are taken at the end. The accuracy of each

.I         reading is taken to be 0.3%.         The error is all random, with no systematic component. Table G.10 shows the resultant sensitivitv of LC?tUF fcr static
~!

storage of f rom 1000 to 4000 kilograms of plutonium. TABLE B.10 Measurement Uncertainty Evaluation for a Static Plutonium Nitrate Storage Facility 1 2 3 4 Number of Bar l -

                       ~
                           .h) 1, . <

s kdj

i

! r.
   .t.

T. , i, I

;d-
'r:                                  TABLE B.li Estimates of Random and Systematic Error for Plutoniun Conversion Facility Accountability Measurements Relatt ve Percent Standard Deviation Material Balance Component               of a Sinile Measurement i'                                                                                    Sampling       Analytical Volume Plutonium Nitrate Input
      ..                                Random                                  0.3                  0.1           0.2 Short-Term Systematic                   0.1                  0.1           0.1 Long Term Systematic                    0.02                 0.02          0.02 7
     >                               Plutonium 0xide Product
   .c                                                                                                0.1           0.25
    .}                                  Randou               i                  0.070                                        ,
   ,!                                   Short-Tern Systematic                   0.035                --            0.046    --

I, ,

                        ~

Long-Term Systematic 0.020 0.02 0.02 0i

v:

e Plutontun Waste

r ,

Random 2 3 10

          ,                             Short-Term SysteEJtic                    !                   --           10 I                    I            I j .!                                     Long-Tern Systematic k't             8

• iiii"-

   .g-                                                                                                     .

t it Ah

l. !

    .s             4 bo ae 7, i
    .' tN, DU,ip I}Ipf4
   's.                        !    ,

i'

                                                                    -118-The systematic errors have been divided into short-term and long-term errors.                      In one accounting period both contribute to the variance equally.                     Over many accounting periods, the short-term systematic error is treated as a random variable with a sample size equal to the number of accounting periods.                                                               The 0.02%                     . residual long-term systematic error is not based                                                                    .

on any evaluations but is thought to simulate the behavior of that portion of the systenatic error that persists from one accounting period to the next. The values shown in Table B.14 will be used in the next section to evaluate the ef fect of formal accounting period length on LEMUF. The effect of the measurement control program, simulated by randomizing the short-tern systematic error component between accounting periods,will also be described. The formal accounting procedura assumes that an attempt has been made to minimize the amount of material held up in the proccsc. The wet side of the process is drained and flushed to remove almost all the plutonium held up. The dryer-calciner is cycled several times in an attempt to dislodge plutonium oxide caked on the walls and in the screw. This process requires 4 to 5 days c f downtime. As a result of this operation, the total amount of plutonium in the sys ten is estinated to be 1.5 kg held up in three l places. One half a kilogram in each dryer-calciner and one-half , kg in the plutonium oxide loadout facility equipment. .1 e i

         - -    . - - . _ _ _ _ = -        _ .  .   ... . ..    . . - ..                                                                   .,;., .

7.- w m

Q .wr::. a-:.: =

a '_ :;- +.~.w

                                                                                       ,**~;' :
    ~s              -

j.;, . ;. ,:.7 " ;. -

                                                     -119-An alternate to this procedure is a draindown inventory
       ~

measurement. As was the case with the formal inventory period, the wet end is runout drained and flushed. This operation takes about a day. However, no attempt is made to cleanout the cal-

     .         ciner. In this case, an estimated 4.5 ko                of plutonium is held up in five places.            There is one kilogram in each of the dryer-calciners and one kilogram held up in each of the drum filters.

The remaining one-half kilogram is in the plutonium oxide load-out equipnent. The running inventoryevaluationisbasedheavklyinthe g facility description provided in the SAR on the plutonf u.w i conversion facility at Barnwell. OD Table B.15 was prepared to summarize the inventory measurements required to perform a running inventory. The measurement error uncertainties are based on two factors. The first and most important is the possible variation in holdup which is likely to be experienced during operation. The second is the estimated accuracy of monitoring normal process variances. Unfortunately there are several pieces of equipment which could experience a large holdup variation. The partially full input piping, the filter and the dryer-calciner are the largest sources of process variance. In addition, in each of the above cases, the geometry of the held up material is difficult ( ' Unfortunately, geometry plays an important role in to predict. determining holdup using NDT techniques. Thus, fairly large measurement uncertainties were placed on several process measure-O ments. The next section will apply those measurement accuracies

                                                                                                                              ~

7 .n

                                   ,                   'J.            :.      .% l [ 7._ , '<" {.    [ ~2 . }. N } l ~

r

                                                                 -120-TABLE B.15       Running Inventory Measurement Uncertainties for the Plutenium Nitrate to Oxide Conversion Facility of a 1500 MT/ Year Reprocessing Plant CU"P " "            Estimated Measurement Accuracy ton Inventory                            %

Process No. of Volume Sampling Analytical Component Components ko

                                                                                                                                         ~

Nitrate Preparation Feed Tanks 2 37.5 Random 0.3 0.1 0.2 Feed Tanks 2 37.5 Syste- 0.1 0.1 0.1 matic . Feed preparation 2 0.6 10{,) Tanks Heat Exchangers 2 0.7 10 Pump 1 0.2 10 Feed piping 2 5.1 25 Precipitation Precipitation Tanks 2 0.53 10 Digester Tanks 6 0.53 10 Vacuum Filters 2 3.5 50 Pnwder Preparation cryer-Calciners 2 4.0 50 Vibrsling Screen 2 0.25 50 Screw Feeder 2 0.40 10.0 10(b) Grinder Assemblies 4 1(b) Blender Assemblies 2 16.0 1 (a) If only one error tern presented, that valve indicates the total measurement error used. (b) Assemblies on load cells. O e O'h m em em-meew - e e w ee -damme. *- ~ . . . . . . .. . . ., [ T '. W , L'J, '

                                                                                              -t             Q
                                                                                                             ~~.;n z =
                                                                                                                            .::t4:1re*7.
                            .-        -~. .                                . a: - ..

m::r. m,.,=

   . .     -~

r ~ - ~

                                                                                             . .~$--    ,                ,_

L t

                                                               -121-to estimate the sensitivity of running inventories to detect losses.

3.3.3 Capabili ty of Ma terial 3alance Accountina System - Tiir a Plut. .. um ni tra te-to-Oxide conversion

 '                            factlity Based on the material accountability I

information presented on the last few pages it is now possible i

      '             to describe the capability of material balances performed over i

4 a conversion facility. Four cases will be considered. First. the ef fect of the accounting perit.d length on LEMUF will be evaluated. This will be followed by the results of the drain-down inventory and then runnint-inventory evaluar. ion. The last topic will be the possible inprovement in los's sensitivity which is expected from the improvements in measurement control. The variation at LEMUF with the frequenc.v of the formal In effect material accountina interval is shown in Table 3.16. there is a factor of two uncertainty in the length of the accounting period for a given throughput level. If both con-version eines are operating,1000 kg of plutonium can be ' processed in 10 days. Twenty days would be required if the plant was operated so that it just kept up with the rate of I plutonium production in the separation facility. The formal, dr'aindown and running inventory analyses j havethesameflowmeasurementvariancesbutdifferentinventor I variances. The variance caicalation for a running inventory measurement is shown in Table,S.17. Table B.-18 shows the effect i of the inventory variance on LEMUF. It can be seen that the running-inventory measurement uncertainties ars'too large to be of much value even when the irventory period is bounded o'n one end by an inventory measurement with a low variance.

                                                                                                                   .e              e=-   e       e -
                                    ,,             ,,               .-.-.            ; ;. . - p      ,  . ~ , - . . .

y - n. . .; ..~.y. , . + . ,

               ,                                                   . .; 7 .-           ,
                                                                                                                           - p' ',; . "               , e' h,. [      i,
                                             ~

S^. p { . . . .;.';

                                                             ,          }l. ~._..f:-   : [}<-           ~.}. y. .: -[.'
                                                                                                                                       .['..f' .::   g' }'[ijj':.:. 2

I l9

      .I, l

TABLE B.16 Hessurement Uncertainty Evaluation for the Plutonium Nitrate-to-Oxide Conversion Area in a 1500 MT/ Year LWR Period 6 daysI ') 12 daysI 'I20 daysI 'I l Month _2 Months 6 Months 1 Year Feed 100 300 600 No. of Batches 12 24 40 49 Batch Size-25 kg of pu 5.25 0.105 0.21 0.35 0.429 0.875 2.625 Random SystematicError Variance Error - kO* variance k98 0.270 1.08 3.00 4.502 18.75 168.75 675 W, . .!' Total Variance - Feed - kg' O.375 ~1.29 3.3b - 4.931 'T9.625 171.38 680.2 I

  ] jja q ', ,

Product 173.42 353.92 1061.77 2123.55 No. of Batches 42.47 84.94 141.57 Batch Size-7.058 - kg of Pu 0.819 kg: 0.016 0.033 0.055 0.067 0.1?6 0.409

   ...

• Random Error Variance 0.674 2.808 25.272 101.088 Systematic Error Variance - ko 8 0.041 0.162 0.449 0.195 0.504'- 0.741 2.944 25.681 101.907

  "};              '       ,

Total Variance Product - kos 0.057 a

                         '       Waste No.of Batches                            10          20        32           40        80        240       480        C N

f Batch Size -(0.025) kg of Pu

l. 7x10-5 1.4x10'4 2.2x10'4 2.8x10' 5.6x10'4 .002 0.003 '

Random Error 0.010 0.04 0.364 1.454 0.006 Systematic .0101 6x10'4 0.003 1.457 8 0.003 0.007 0.010 0.041 0.366 j}' % ,3' h . .' Total Variance - Waste - kg .001 783.614 0.433 1.488 3.861 5.682 22.61 197.427 Total Flow variance - ko 8 8

                   '!                                                                                0.375        0.375     0.375     0.375       0.375 inventory Variance - ko                       0.375      0.375                                                783.80 8                      0.808       1.863     4.236        6.057    22.99    197.62 Total variance - kg 1.365     2.050        2.461     4.80     14.06      28.00
                       ;             Sd(MUF) - kg                              0.899                                                 28.12      55.99 i,

LEMUF - kg 1.798 2.7308 4.116 4.923 9.59 if"(! . .j ,, 0.373 0.599 0.455 0.412 0.401 0.384 0.375 F LEMUFx100/ Feed a) Period could be shorter by a f actor of two if both conversion lines running at design capacity. . .. ); . s 4

          . . . . i e          &
  $}                *SV
  ;;jy
                                   --e .

O h m

l 1

                                          -123-TABLE B.17      Running Inventory Heasurer.ent Uncertainty Evaluation for the Plutonium Nitrate to 0xide Conversion Area in a 1500 MT/ Year Reprocessing Plant Tota Component             Estimated                     Mea s t.reuent Plutonium             Heasurement Error Process                No. Of          Inventory             Accuracy                      Variance Component              Components      ko                    "
                                                                      .                             kq2
        , Nitrate Prepa-ration Feed tanks              2                37.5             Random 0.37                            0.039 Feed tanks              2                37.5             syste- 0.17                            0.017 matic Feed prepa-             2                   0.6                           10                     0.007 ra tio n tanks Heat Exchangers         2                   0.7                           10                     0.009 Pump                     1                   0.2                           10                     0.000 Feed Piping             2                    5.1                          25                       3.251 3.324 I       Preci pi ta ti on Precipitation            2                   0.53                          10                      0.005 tanks Digester Tanks            6                  0.53                          10                      0.017 Vacuum Filters            2                  3.5                           50'                     6.125 6.14S Powder Preparation Dryer-Calciners           2                  4.0                           50                      8.000 Vibrating Screen          2                  0.25                          50                      0.031 Screw Feeder             2                  0.40                          10(3) 0.003 Grindar Assemblies       4                 10.0                               1(c) 0.040 Blender Assemblies       2                 16.0                               1                   0.051 8.125 '

TOTALS 180.44 17.597 a) Assemblies on load cells. N W - e

                                                                        "'y                    *E_,  **F'*          *

• M%vo

                                                                                              .n             .

a * * - --u. a. wq , . , ...g '_ ". , ,,

e iffi

• TABLE B.18 Comparison of LEMUF Sensitivity to Frequency and the Type of Physical Inventory Performed tu Obtain the Material Balance for a PlutJnium Conversion Facility LEMUF Fornal Accounting Inventory Drairoown Running Measurement on One Side Inventory Inventory of the Accountino Period On on Both Sides Both Sides Draindown Running of the of the Fornal on on on Accounting Accounting

 "'          Accounting                                                                            Period l  Period                             Second Side Second Side   Second Side   Period 8.53          3.20     11.94 6 days la)                           1.80           2.59 2.73           3.31        8.78          3.80     12.11 c       t   12 day [J) 4.12           4.52        9.30          4.89     12.50 20 days (J) 4.92           5.27        9.68           5.59     12.79 1 Month                                                                                          m]

9.59 9.76 12.71 9.95 15.21

g. j 2 Months if a) Could be shorter by a factor of two if both conversion lines operating.

flhk sjp

                    . g .                                                        $      0 e

                                                - 125-The tradeoff between draindown .nd fornal i nventory neasure-ment sensitivities is much smaller.           Indeed, sunnlementine, the f a nal accounting records with intermediate draindown inventory analyses would appear to be worthwhile.           Especially, since npera-
     . tional requirements may requ're draindown inventories to separate batches for various customers.

The effect of the measurement control progran on the long-term sensitivi ty of the plant to detect small continuous or semi-continuous losses is shown in Table B.18. Large gains in sensitivity as neasured by the decrease in the term (LEMUF x 100/ Feed) can be realized fron the measurement control orogram. D.4 CAPAGILITY OF HATERIAL CALANCE ACCOUNTIfic. SYSTEMS F0k P L UT ClI U.*4 IN A 200 MT/ YEAR MIXED OXIDE LIJR FUEL FABRICATION PLANT The description of a 200 flT/ year mixed-oxide LWR fuel fabrication plant is based on a model develooed by E. Sain, et al.IIII The naterial balance data taken from this report is shown in Fiaure 3.3. It should be recognized that one 1500 MT/ year reprocessing plant separates enough plutonium for a mixed oxide fabrication plant having a capacity of approximately 400 tit / year. This factor should be included wherever a comparison of plant types is made. The organization of tnis section will parallel that of previous sections. Following a description of the important material accountinc features of the facility, subsections will l be devoted to an evaluation of neasurement uncertainties and finally, a sunnary of material accounting capabilities. S. A 44 g 4 s 4 e t k M '4 e $ 4 f - *,#h

9t e f u, . TABLE B.19 Effect of the ibasurement Control Program on the Long-Tenn Measurement Errors for the Plutonium Nitrate-to-Oxide Conversion Area for a 1500 MT/ Year LWR Reprocessing Plant 2 Months 6 Menthe 1iear  ? Vrart Elapsed Time from Initiation of Control Program 1 Yeart I Nucller of Inventory Periods 1 3 6 a Yeart E Yeart E Venet A Yeart la Venri 12 18 24 30 36 I 44 60

                      ?                                    Feed - 25 k9 Batches                                                                           Cumulative Measurement Error Jarlances No. of Satches/ Inventory i
            .3                                                    Period - 100
                                                                                                           .875
       '{f. p* i,                                             5'. ortdos Ree.          IcreError Vartance - 69:

Systematic 2.625 5.25 10.5 15.75 18 54 24.875 26.25 31.5 42 52.5 IC8 216 5 Error Vartance kg 8 Long Term Systematic 324 432 540 648 864 1080

                                                                                                           .75      6.75     27       108       243 g

Error variance kg 8 432 675 972 1728  ??00 L, 10tal Variance Feed kg: 19.625 63.375 140.25 334.5 582.75 888.875 1241.25 1651.5 2634 3832.5 Product . 7.058 kg 8atches No of Batches / Inventory .

           ';-. * -                                               Perted           353.92 Random Error Verlance -kg 8                 .336      0.409     0.819 Snort ferm Systematic                                                       1.638     2.456        3.275      4.094       4.913       6.550 2.059       6.177    12.35     24.75      31.06                                                       8.188 Error Verlance - kg8                                                                         49.42      61.77       74.13       9A.84     123.55 tong fera Systematic                       0.748      6.739 s

26.96 107.82 t, Error Vartance kg Total Vartance - Product kg W W 8 242.51 431.30 673.90 970.42 1725.19 2695.51 g g ! !' 8 W ITT.17 ITCTT TIT ~T5 W 16 TITC3T'" N a Waste 0.025 kg Satches a ' g'i No. of Batches / Inventory Perlod . 80

                       '                                    Rand:e Error Yartance kg 8                  0.001      0.002      0.003     0.007 Short 1ers Systematic                       0.040                                      0.010       0.014      0.017       0.020       0.027      0.034
                      .                                                                                            0.12       0.24      0.48       0.72        0.96 Error Verlance kg a                                                                                      1.20        1.44        1.92       2.40 Long Tere Systematic                        0.001      0.011 i*

Error variance kg8 _0.043- 0.173 0.389 0.691 1.08 1.55 2.765 4.37 T Total Vartance - Waste kg a 0.042 0.133 0.286 0.66 6 1.119 1.665 2.297 3.01 4.712 6.154

         !.                '>                             TotalFlowVariance-kg8 Invento/y variance . kg 22.61      76.833    180.67    469.33    866.00      1314.54    1983.3     2703.97     4469.3 0.375                                                                                             6666.5 Total Measurement Error                     22.99      77.21     181.05    469.7     866.38 variance                                                                                        1374.96    1983.7     2704.4      4469.7      6666.9
                                                         $d(MUF) . kg                                  4.795      8.79 LEMUF . ke                                                          13.45     21.67     29.43        37.08      44.54      52.00       66.85 9.59      17.57      26.90                                                                          81.65
           ..),p,                                       LEMUFal00 Reed                                 0.384      0.234      0.179 43.34 0.144 58.86              5    89.07     104.00      133.71      163.30 s

0.131 74.]23 0.1 0.119 0.116 0.111 0.109 k%

                    . s .<

h,

       &{h ti
                                                                                   .

• 9 . .

                                         .e m        e         *

                .s
             *f
                                                                                                             -127-
                          ...   . . ,o..                 .....                                                          , ...
                            .,i                          ._.                  l                      le.                   . . .                     l
                                            ....,4.....a.....                                               ...4..

10.... ... . . . . . . .. ... l le ...._,3. _... .. ... ,,....... l

                                            ....4........
                             -lje ~ , - ~                                     l. .:: .+              le                , - -

l

         .                                  . .. . , s                                                   ... . .. , s
                          .- ,s ... ~. ._. ....
                                , , . . . . .                       ...       l... .. ,              le. ...,. ..       .                            I                     le           ..

l..,_, . a~~. "". .  % I .i 3, ... .. .. .. ... l . . .. .~. j e. . . . . l

                              ...             . . . . . ~ , . . . .                                          . ..           3.                 .           . , - .

le . .... i

                                                                                                                        ,t..                                                   e ... ..                           :;~ .,,
                                                                                        . ,..e i..~,                                     ....

le .. . .. . .. - . . . . . . I

m. , 4 . . ..

e i

                                                                                                         -...,4...

a. -

                                                                                                                 . . . .. . . ..                                                                    =~

lO . . . ...... l

                                                                                                        -.-,4--.                                        - .,
                                                                                            . ,... e                  . . . -                           %
                                                                                          ...                                                           i   - -           z.
                                                                                                            ...~.             .
                                                                                   .....,.le  ..          . . . .                  ......   ._ l                                                    . . ~ ,
                                                                                                         .....,4,,......

(  :::.-a., je . . ...~ 1 t. 1 ,.,,..,4. . . .- ... e . ..

                          -l::* .",,.-",; "-~

le ... , 4 ,o i _ . .. ..

                                                                                                                                                                     ~. ,
                        . . . .. .. .. . .           .,                                                                                                r        ...
                        ......          . ,                                                        Is i      n.....a--...-...-...l 4 , - . .
                                      ==                . ,                                        le..~ =...=. -                                      l
           \

e FIGURE B.3 Average Material Flow Per 8-Hour Shift for M02 Fuel Plant l \ .

                                                                                                                                                                                                                                ~
                                                                                                                                                                                     ._           _                +_ . . .

n' *d'b

• m , OB- ^ -

9- 4 m

• 44 M 9 a e

, e--

                                                  -128-B.4.1    Description .>t      Imoortant Haterial Accounting Design Features The fuel fabrication plant described                      .

in Reference (11) incorporates many of the design features suggested in regulatory guides 2-14) for isproved materials accounting. The following paragraphs highlight those features. . Regulatory Guide 5.8 discusses design considerations to minimize holdup in drying and fluidized-bed operations. The - 2 input powder, the M02blended p wder and the M0 recovered Pu0 2 scrap are all stored in fluidized bed silos. Based on experi-ence with blowback filters.U6 a Pu0 2 input storage silo may hold up from 500 grams to 3.5 kg of plutonium. A value of 1.3 kg was used in this analysis. The valve at the bottom of the silo may also hold up some Pu02 powder. However, the entire assembly including the outlet valve is on a weigh cell. Thus, the holdup level car. be nonitored. Regulatory Guide 5.8 also discusses the problem of caking in fluidized beds. Both the Pu02 and mixed oxide are very stable chemically and by closely controlling the noisture, some-thing essential for criticality safety, there should be no caking problem. Once again, since the entire silo is on a weigh cell, material held up can be weighed very accurately. Regulatory Guide 5.25 discusses design criterion for wet-process operations. Except for the lab, there are no wet processes. Thus, the design criteria in 5.25 are automatically

• met.

Regulatory Guide 5.26 discusses naterial balance areas and item control areas. The proposed plant contains a process computer that monitors the tare weights, gross weights and

                                                                                 --kR        - - .
                                                                   %_ .          m.      #
                                                                                           .y-
 .a  3,p-      ==       .::y=;:.;        3.y .         3.;         3            my.      3 g_

                                               -129-locations of all in-process pellet trays. Most tra.,sfers are computer controlled. Scrap transf ers are done remotely. Thus transfers between MBA's or ICA's cannot occur without beinn monitored by the process computer. Even though some areas do not physically separate material in difficult bala 'e areas, such areas are in remote canyons and there is no way to transfer materials from one area to another.

Regulatory Guide 5.42 discusses design criterion for dry-process operations. The facility places heavy reliance on oneunatic transfers. Storage hoppers have conical-shaped bottoms and Pu0 2 is non-re ctive with the materials of con-struction of process equipment. Moisture content, a potential problem, is monitored and 'must be controlled for criticality safety. Based on the regulatory guide, the above design characteris tics are highly desirable. The only apparent deficiency is the i

 ~

inability to measure the amount of material held up as feed to the slug, granulator and pellet machines. These feeding hoppers muct be run out, a orocess taking 3 to 6 hours, before an accurate inventory estimate can be nade. The above paragraphs briefly describe how the design l described in Reference (11) meets or exceeds most of the regulatory e design criteria applicable to fuel fabrication facilities. Sub-sequent sections will discuss measurement uncertainties and

      ,         the expected material accounting capability of the facility.

l ,

                                                                                    ,- r. n _
          .' :.    ._  _            -          _. 2..  ,

u .

,e

                                                    -130-B.4.2   Measurement Uncertainties for Mixed Oxide fabrication Plant Plutonium Accountabilib fle a s u reme n t s The measurenent uncertainties associated with what are considered to be the best formal accountino measure-ments will be described in this section.               The uncertainties                      .

associated with running-inventory type accounting systens has been discussed earlier, urider the topic Raal Time Material Control - and Accounting Systens. As was the case with the reprocessinn plant, there are many years of experience with naterial accountino in a mixed oxide fabrication plant. Perfo*mance data from this country and abroad is available(2,3,4,6,7) Tab ~le B.20 reoresents a comnosite summary of the estimated present capability of fabrication plant measurenent systems. There have been no studies to date which break the systematic error term up into long-term and short-term components. As e result, a long-term systematic component of 0.2% was selected to represent the fraction of the systematic error that might persist for many years. The short-term component is simply the difference between the systematic error component es tima ted f rom past studies and the lonn-tern systematic error tern. Since the long- and short-term conoonents must be squared and then added to get the total systematic error, some of the short- term errors are presented as two place decinals. This was done solely to make the total syste-ma tic error come out as a single digit number. l

                                                                          ~
                                                                                      ./GM'
                                                                                       . r.        3:
         ~

_ _~~  :- . .... - :~L - 2,= nk' ..

                                                                                                  - $En
                                                                                                   ~
                                         .                    :v                  ::          x    ca=r:
    ; ;~~ /

_ }} . -(? {' 5- - .:L} ; ~

                                                                                              /   'l     ~*

                                            -131-TABLE B.20   Es timates of Random and Systematic Error for Fabrication Plant Plutonium Accountability Measurements Relative Percent Standard Deviation Material Balance Component                '

of A Single Measurement Weiching Sampling Analytical Input Plutonium 0xide Random 0.07 0.10 0.25

   ,        Short-Term Systematic                      0.035      --

0.046 Long-Term Systematic 0.020 0.02 0.02 Products Plutonium Output (Finished Pellets) Random 0.07 0.10 0.30 Short-Term Systematic 0.035 -- 0.046 Long-Term Systematic 0.020 0.02 0.020 Plutonium Waste Random 2.0 3.0 10.0 Short-Term Systenatic -- -- 10.0 Long-Term Systematic 1.0 1.0 1.0 Dirty Scrap Random 0.1 5.0 5.0 Short-Term Sys tema tic -- -- 5.0 Long-Term Systematic 0.1 1.0 1.0 In-process Storage Mixed Oxide Powder Blends (inc. recovered clean scrap) Random 0.07 0.10 0.30 Short-Term Sys tema tic 0.035 -- 0.08 Long-Term Systematic 0.020 0.02 0.02

  ,       Mixed Oxide Scrap Can i

t Random ~ 0.07 0.20 0.30 Short-Term Systematic 0.035 -- 0.08 Long-Term Systematic 0.02 0.02 0.02 l Green Pellet Store l Randon 0.07 0.15 0.30 l Short-Tern Systematic 0.035 -- 0.08 e Long-Term Systematic 0.02 0.02 0.02 Sintered Pellet Store Randon 0.07 0.10 0.30 l

                                                                                                    ~~

e W g g ,, , w g , e O +,+s* we

                                                   -132-B.4.3     Capability of Material Balance Accounting
                                  ~

Systems tor Plutonium in a 200 MT/ Year Mixed Oxide Fuel Fabrication Facility From 'the s tandpoint of material account-ing, major emphasis has been placed on overall plant performance. , When discussing overall performance, measurement errors associated with internal transfers are not included. Intermediate materials. . present as in-process inventories do enter into the inventory term Thus, in-process inventories, i f they are present in well characterized forms, do not seriously degrade the sensitivity of the accounting system. Tables B.21 and B.22 have been prepared to summarize the estimated contribution of in-process material to the overall plant accounting system variance. Table B.21 produces a tabulation of the measurement variance for material held up in the process vessels. The holdup-level estinates have been based on a facility decommissioning study.(16) Table B.22 provides a tabulation of the variance 'or material held up as well . characterized in-process inventories. These inventory variances are combined with the flow variance terms for several accounting periods to obtain the sensitivity of LEMUF to inventory level and inventory frequency. Table B.23 shows the results of combining the high inventory-level variances with the flow variances for several accounting

• periods. Table 8.24 comoares the value of LEMilF obtained for the high inventory case with a case where the well-characterized in-process inventory level is small. The gain in sensitivity of loss detection is very small.

m y Ni ug k.

     --.               --.                . - . , - .       . ,n,           .=_
                                                                              .       ,. w         . n. .
                                              ;-y *
    . = 7:      . ;-          a.; ;.:,2

a';g.*,.  :

Q.

MW '

G ~~' f

 ,e
                                                -133-TABLE  3.21  Inventory Measurement Uncertainty Evaluation for
         -"~ ~

Material lleid Up in Process Equipment in a 200 MT/ Year Mixed 0xide fuel Fabrication facili ty Plutonium Total Process No. Of Holdup Measurenent Measurement Component ComDonents per Component Accuracy Error Variance 2 kg's 7. ko Pu0, Powder 3 0.2 50 0.03 Unioading and Transfer Pu0 Storage 3 1.3 0.29 4.2x10-5 Sil$s Blender 1 1.0 50 0.25 Pu0,d fee System M0 Storage 9 0.05 0.29 1.9x10-7 Sflos M0, Blender 1 0.1 50 2.5x10-3 Fbed Systen Blend, Mill 1 0.05 50 6.25x10~4 Slug, Granulate, 3 0.05 50 1.875x10-3 Press . Grind 2 0.05 50 1.25x10-3 Pellet Inspec- 1 0.1 50 2.5x10-3 tion Lab 36 0.05 59 0.0225 Defect Rod 1 01_  :> 0 2.5x10~3 Unioading Clean Scrap 4 0.06 50 3.6x10-3 i Recovery s Misc. Waste 11 0.05 50 6.8x10~3

       . Treatment Clean Scrap          3             0.05                0.29        1.9x10" Storage TOTALS                            9.25                            0.3187 i

1 s - j - 4.. p.y

l

                                                 -134-I TABLE B.2;2    Beginning Plus Endinn laventory Measurement Uncer-                              l tainty Evaluation for Well-Measured in-Prncess                              -

l Inventories in a 200 MT/ Year Mixed 0xide Fuel Fabrication Facility , 1 Total Total Plutonium Random Systematic Total

• Process Content Error Error Error Inven tory No. of Per Item Variance Variance Variance Item Items (kg of Pu) kg# kg# ko? ,

Pu0 Powder 2 150 0.348 0.020 0.368 S los (a) pug Powder 7 1 150 0.174 0 0.174 SiTos 2 Powder Silos 7.94 M0 5 0.0047 0.0013 0.0060 (agleanScrapSilos 2 25.41 0.0100 0.0012 0.0112 Clean Scrap Silo 1 25.41 0.0050 0 0.0050 Green Pellet 29 0.3176 3.0x10-5 7.0x10-5 -(0.0001)2 trays (b) Sintered Pellet ' O.3176 7.0x10 -5 3.8x10~4 (0.0004)2 trays (b) Finished Pellet 444 -4 0.31/6 4.7x10 1.7x10'2 (0.0172)2 trays (b) Clean Scrap 35 0.7058 2.0x10'4 5.0x10-4 cans (0.0007)2 Subtotal - if some silos inactive for accountinn oeriod 0.4280 TOTAL 762.8 0.6020 a) Not appiicable for accounting periods where when one silo remains inactive for accountino period. b) Different batches are assumed present at beginning ar d end of inventory period. Thus, total variance must be muitiplied by two. 9 4% - .

                                                                                   . MT:           *C-Y'                                        f
      ~
                                     = ,___.       . . - . E. -  ---     -
                                                                                   .y

_ = .- -- 1.=_ =g = gw =x e =r

 -                     =      =                                                                                          .

o  ;,.. TABLE D.23 Measurement Uncertainty Evaluation for a 200 "T' Year Mixed Oxide Fuci Fabrication Facility Len9th of '4aterial Accounting Peried 1 Days 15 Days 1 '!onth 2 Months 6 Months 12 'tonths

                                                                                                     **casurement Error Variances t          Feed - 7.053 k9's of Pu                                                                                                   1012 Humber of Batches                                                  21.25    42.5      84.3          168.7    506 Random Error Variance - kg'                                         O.0082   0.0164     0.0325        0.0650    0.195      0.390 Systematic Error Variance - kg 8                                    0.0101   0.0405     0.1594        0.6378    5.740    22.958 Total Variance - Feed - k9                                      8 0.0183   0.0569     0.1919        0.7028    5.935    23.348 Product - 0.3176 kg's Number of Batches                                                 466.62   933.2    1852           3703.7     11.111     22.222 Random Error Variance - kg                                          0.0005   0.0010     0.0020        0.004     0.012      0.024 Sys tema tic Error Variance - kg'                                   O 0184   0.0738     0.2906        1.162   10.461     41.844
  ,             Total Variance - Product - kq-                                      0.0189   0.0748     0.2926        1.166   10.473     41.868      ,

w . Oirty Scrap - 0.10G kg T l' umber of Batches 8.39 16.3 33.2 66.4 199 398 Random Error Variance - kg 8 0.0005 0.0009 0.0019 0.004 0.011 0.022

  ,             Systematic Error Variance - ko'                                     O.0020   0.0079     0.0309        0.124      1.113     4.453 Total Variance - D. Scrap kg'                                       O.0025   0.0088     0.0338        0.127      1.124     4.475 Misc. Waste 0.1882 kg Number of Batches                                                   4.72     9.45      18.7          37.4    112.1      224.3 Random Error Variance - kg'                                         O.006    0.0013     0.0025        0.005     0.015      0.030 Sys tema tic Er or Variance - kg'                                   O.0020   0.0079     0.0309        0.124      1.113     4.453 Total Variance Maste kg'                                            O.0026   0.0092     0.0334        0.129      1.128     4.483
    .        Total Variance - Flow kg                                               0.042    0.150      0.551         2.125    18.661    74.174
        -    Total Variance - Inventory kg 8                                        1.239 Total Variance - kg                                                    1.281    1.389       1.790        3.364    19.90     75.413 34 i'          Sd(*lUF) - kg                                                          1.132    1.178       1.338        1.834      4.46      8.68 LEMUF - kg                                                             2.26     2.36       2.68          3.67       8.92     17.4 LEMUF*100/ceed                                                         1.50     0.79       0.45          0.31       0.25      0.24 W: 2 1

m W \ i I 1

                                         -136-TABLE 0.24  Comparison of LEMUF Sensitivity to Frequency and Anount                                                 l of-Well Measured in-Process Material Included in the                                                    1 Material Balance for a 200 NT/ Year Fuel Fabrication Flart                                              '

l No In-process Large Amount Materials other of in-Process Material than Quantities (762 kg's of Pu) . Accounting Held up in in Addition to Quantities Period Process Eouipment Held up in Process Equ;oment kg kg i Week 1.649 2.264 2 Weeks 1.775 2.357 I Honth 2.180 2.676  ! 2 Honths 3.324 3.668 i l I W 1 l 1

                                                                                                                   ~

l t

                                                                                                   .Q&      --e-l awe = * * - " * * *

2. e.

se '

                                                                                                "-   P.  **
                                                                                                          ,** y *.

9.- r .43. ..

                                     % 9=.                   . -z r9                       h_, :-   15:

, y-137-The same data used to develop the results shown in Table B.25 can also be used to show the effect of randomizino the short-term systematic error from one accounting period to the next. It is believed that this randomization will be a result of the measurement control program.I9I As shown in Table B.24, randomizing a component of the systenatic error results in a si;nificant reduction in the rate of LEMUF/ Feed. As discussed in the introduction to section 3, this ratio is a measure of the detection sensitivity of the plant anainst small con-tinuous or semicontinuous losses of naterial. The previous paranraphs discussed the behavior of the overall plant balance. Based on the error terns in Table G.20 it can be seen that intermediate process flows can be measured with essentially the same accuracy as the incut and final product neasurements. This means that material balances taken over snali pieces of the process m3y be tir.ely, sensitive lots detectors. Table G.26 shows the results of a material balanc? taken over the plutonium oxide powder storace area. Since the input is weighed and anal,vzed as part of the plant balance, this term is identi. cal to the input term in Table 0.23 The storage silos are on weigh cells and can be measured accurately. The outnut is bat:hed into the Pu02 blender input hopper which is also on a weigh rell. Thus, all the flows and inventories can be determined and monitored in an essentially continuous nanner. Based on the resul ts , shown in Table B.26,

                                                                                                                             ~

G ep $ 4 g e,4 -,g e w % ., _w *8- $ = pm =t- - r- es - w W w

i I l

                                                                                                                        -138-                                                                                                                                                                                                  ,

1 0 . I l

• P. . ee e .-

b e e

                                                                                                          #=

c=*= a 3 e, - ~cm e o.. p* c**, **,P.4.* we i e., %,. si es -i. e, e O.

                                                          .                     asL' .                    ee ~ ~                                       ev. e,ese.                    e                           e .P.
                                                                                                                                                                                                                                                      #.           .         e e                                        .
                              > O

• O* *
• e e o e .

C*=* **I O  % **  % ** ?. e P* E em C. dp . ** .3 e c' h* e, .: .'s .,,

                                                                                                                                                                                                                                                                .           -s O                                     N                                          en        e          ,                                #             e w         w
                              ==    :                                           -e9                                      es                                                                                                                           am        -

9 <s

• ae ** == n bo er <* * * ,

e,- e en r* e ee, ,e,b,,c,, e,..  %. gp y g b N M

                                                                                                           .h.*  PS      .e.'     ),e                   e,u.m ,m,,O,           t                              ,                            ,          ,g
                                                                                                                                                                                                                                                         . .,g. ., . . .

ae* go * *mp e s- . e.8 g * = ,e . . . . . W C* C' O. C g  % e *

• O ** h T"* O en se ** en se ' t* A

                               >     4e                   *9        *8          **

e* em <==l m ee ne en {, gn

                       , ei          ,

8'e 8 m'*ee* e m' f a.w.

           ,           b
                       '                                                                                                                                                                                                                    e                                           e em            ee         m                                e                                        - es ec                      n          es                   - e= O 3                                 e,                                                                                   te,a                 e .v

r. et , e- n O

era o e e ~e O ce est sw 89 en W *e w ** en b r1 ,re. * = e q A, b g C e9 eCe e. e sie e a). ce. a-. e.n e i

                                                                                                                                                                                                                   +              -

M e h* e M ene O* gg es & no e D eft e - OM r** OMe EP M

• e te G C > r=== De e D e* M a== i.9 a==56re e* Se e't y , * *

           ==           O                      b                     a=*         eis                                                                                                                                                                   en to es                                       De        re                                             r 8""   be      edI             e E-6 U        -

GM * '*

                                                                                                                                                                                                                 -e             er           e.                                         C em       en         N                   -e            <>                                                                    %         O                     he-
           >= 6         C        eg                                                                         .e.- e                   e                                 e                                         ce            es.

e. ert es e <e

                                                                                                                                                            =e         m                                         - ee           a            e-        O          ce en - -                                                  *
                                                               **o*

8 6 ***

                                                                                                            - a>

O *' -" -**e * * *'e'* -m*

• O ie * * - *
• O* C -O e e.e en

                                                                                     -

• ee* - *ce* *
• C O ** 6 O

                                                           ** * **                                           ***           *         *                    **

• r**

Ow * * '9 8* ew - #9 Mg - C - en am g=

                                                                                                                           ,..                                                                                                                          est es      ep (           et P3   -                       E                                      e
             &U                                 e g .e.       .C.                    b M                                           ==

3 e se O >= == ee C M L

• e at == en en en eie == ee ec me up w e ==

g 5 O k e e ee es e cum e. e e e=

h. . e. e.

N

e. p e. . ,

O b e 6

                                                            #.        N.           P.

N O. p. e. . D. e. c. l

                                                                                                                                                                                                                                             *= **                 *e en - O O Ia-      *         *ez                                              **                        O    #9        en                             C' ee e                   p+                           O es e                     --         aa         =a           -M

>Fw em em= 6-D G' -

= e en se e O en en e9 . - - L g e == 3 a. m O .m en O es 5 em ee C a= . O em ==. ee. e=t 4 .g b @ O. pi. 32 e. e. . es.==. Eh e 3 em. e. a=. O. en. e. . * = . e a- As ae c '8" e a e ha e en c9 O ne er* O s= ay .O 6 -N 8** OO =,=N ae > >= s-e Om- e G em -en 74, 6 M *C C O t es e-OX t W trJa Oe ec) = U- e, O ce cPe m. e a=- ew .tv es ee O c9 e3 @O * * * , O Sh e9 em eis eu se e >=.yen . coeget. ,= 7 bi Ca dP =* @ O e9 *= $* O rt a- g e, en. . e e=.= -es. O. g=. * . Ng mo e

• M *er'e== s.t .

#9 et g e en e O- am e O- N b* D a== s h e'un - > se N r9 Pe NO no e. I CJ u-b %s Nm e p om et A MOO U O e we P5 ep *= #9 h gr en e. O N Nb m.= ew e% e. e r0 ed es *9 r3 %p6 e cm la 4 b' e em ce op ==. . en p O. es. en. r e. c. e. e . s. h. . e=. =Jn.a . O. w.n e weQ yoe. , s w e N e De e L. == CO O - OO O{ ~ m ee 8% N ** U 2 4 9 '#'** * * ** - ** af* *a b sa * * * ' '

• 98= 0 O%

~ C s'* 4e ** 86 up "O "'t .s. - p= ~ "89 C "er . "n. e "e. "*w". .e . O r'9 e e en. . O. em. se. . OO O* C OO O O es en et e O = a= O ce gy M 4M E I Vb

• GO

 % b p %b L .e es e ww e en on en m em s en e e er O P9 e- er oees. e U e e* OO .3 - N OO O-c.e .ew . - ew e9 ee, en m e .O.e 3 O., O. p .e. O se s-O==- h* * * . . as. e. . Ol.O. Na ggg C , e

• a O- e =n OO O= OO --O CD. o. g N x .

O Os . mis ap. "sp d J EP Eb O s an, g a t. e b esyn .gp b b ,,on O e- e E[m e , O

p. . .e se

,.up e w e G. p b u e#* ab b O b ei o cm sg es as b  % e e e # st O b 8' h 88"8 b JS b W.  % &# O en b . 4. e 3 mb Ed' #8 W h G' b ens GP y W eO Sp p E* *= E* O eA b .e . O C o ~ = O ou ~ se -e w- o e. e -a ee < a a b -e.am. e. .O .o .u. .e C 68 v .aoOb c e & e.

• 88 u b

*C C *d ** 9 *B R Sp se b GP  %# 6 08 e*** 6 4 ek e em C to ep C to g g e ed em e& e9 D Gr b 9 9 *8 O

• GB M #
• 9E 6 *8 9

O> w= C Eb eC et @ om b e tm 3 - E b ge g g er e e OC @ l'8 ** E.* Gr E

• C 4 b" b.. E en e

(** %# t'* te a.og ene g Sb o. e.* tm gg es v en es .e C g e.* w GPW ****Pt "" C# e *D

• gye

 %  % 2. D et C %C Sp *e 8 **d U" # d '%"" *3***.e**4 8"% 9 ed es e e as g, b W  % 4 eve se u @% == rume := uft C e8 C b . *

• a .se .s en g .o u ,

t' et, a- O GP e es e 4 e, og e. b W 8 @t . en e d O a g b qpe == w g om ese M 3 , e em e u an e C  %* C 66e w e O em Gp O 7C Q ga g gyy C C ewe k g as g b E bas .g gig g ei. bg en == g sine ga g=Cg ag O es9 9 MW ebT& km @ O ep b b>ghk .wmbgg g- W ebb bEbw E w"db* b b ogog g g C *** O D # eDwe > g Ow b b>E>b g g Lw g as g%s gs C b C e .wg g est to C e. C g e-- 4 ,9-C O. S O pe b 4. ium e O ene gp em bgb> b JB e D* 9 G8 93" *"6 E3 O t= g er g mm g gg , em em E OOO *E **W"* ** E - ee - g ag , ,,  % b== O ed b e* b se e 33 O. F O em O ta * * *

• b O *# b 'P b" 84" b 3 e8 b b am 3m p p ,,,,,g e8 h b. O e8 h b**

O e O GP W b bob g O ** b 9

• O t' O b g cmoe w O aC O ne a ==4** u o es 93 h e. EP b
• bCOD CDW 9 ---

*K O S & C. ,O w O CwW O -O == 0 0 -e.b C. ,O D CO D W . ..g ,3 3 ,

9. ,0 . . - EW J * #F MW J > www .gg 9

b uO ,O M@ J > *bE

• O E
• b E p. ,O,, ,O, ,O,,, w , ,J,,

& O - ' = ' ' * . * * - = . ' . . .. , , _ , ..,a-e~ . .e m . -e.g # .J a e+= 'O 3 N. # ^Y**'"*

• 7 ".' ...w s

.e,. y - A .'._b,  % , f* l .,3 e. ., q h e.. s - ,f. wm w.. -.. .~ *w_. c .* . ' . .* .e.,a,,,.Qa, w e.-~gn ,. , . .,. w #Q ,. . :. .a v

v. - w. a=

c.sm, ewe , _ . . - . . . . ~,, i.ee ,

• G * ,

4 1 k Measurement Uncertainty Evaluation for the Pu02 Power TABLE B.16- , Storage Area in a 200 MT/ Year Fuel Fabrication Plant T1me Peried .htg JB,u r.,g h Eggk, 2,,jfte13 1 unnt5 Mf,LM i* Flom Terms Feed Component No. of Batches 1 1.012 3.036 2.25 42.5 84.3 168.6 i Satch Stae (kg or Pu) 7.058 , Randon Error variance 3.86s10~8 3.90a10** 1.17:10*3 0.008 0.016 0.033 0.065 gga i 2.2410-5 2.30:10*3 2.06:19*4 0.010 0.040 0.159 0.678 i Systematic Error l Variance kg 8 . Total Error Vart. -- 4.13 10*' 1.38:10*3 0.018 0.357 0.192 0.703  ! ante hg a ,i i Product *cmponent  ! No of Batches I 1 3 21 42 83.3 166.6 Satch Site (kg of Pu) 7.142 . Random Error Variance 3.86x10'4 3.86x10~4 1.18:10*3 0.009 0.017 0.034 0.066 8 l k9 Systematic Error 2.24 10-52 .24:10-5 2.06:10~4 0.010 U.040 0.159 0.638 I Verlance kg 8 .. I Total Error -- 4.08:10** 1.39 10*3 0.019 0.057 0.193 0.704 f ',

• I Vartance kg 8'

' Flow Total -- 8.21 10** 2.77x10*3 0.037 0.114 0.384 1.407

i. so I Inventory Terms i PuC2 5torage Silos No. of AClive $llos -- -- 2 2 3 3 3

, Beginning Inventory kg -- (120. 30) (120, 30) (150. 0) (150. 150. 0) (150,150.0) (150. 150. 0) l (511o 1. $11o 2 Sito 3) (150. O. 150) Ending Inventory kg -- (113. 37) (100. 50) (0. 150) (150. D. 150) (150. O. 150) l (Silo 1. Silo 2 ) Silo 3) 0.6965 0.6966 Random Error Vartance -- .2278 .2152 0.3483 0.6966 j ItT Systematic Error 4.58:10-5 4.58:10-5 1..r7:10-4 0.0206 0.0206 0.0206 0.0206

• Variance Total Error variance -- .2279 .2153 0.3689 0.7172 0.7172 0.7172 8

Holdup inventory 2 2 2 3 3 3 No. of Location -- Amt/ Loc (kg) -- 0.2 0.2 0.2 0.2 0.2 0.2 ' BI Random Error , f Vartance kg 8 -- 0.02 0.02 0.02 0.03 0.03 0.03 ) El Randon Error -- 0.02 0.02 0.02 0.03 0.03 0.03 Vartance kg8 - Total Error Variance kg 8-- 0.04 0.04 0.04 0.06 0.06 0.06 Total Inventory Error -- 0.2679 0.2553 0.4089 .7772 .7772 .7772 { Variance kg 8 Total Error Variance kg 8 -- 0.269 0.258 450 .891 1.861 2.144 l 1.08

• 4%

I 5d(Mur) kg -- 3.52 0.51 0.67 .94 . LEMUF kg -- 1.0 1.0 1.3 1.9 2.2 1. ',  ? LEMUFal00/ feed 3 -- 14.5 4.7 .89 .63 .36 .25 mW l -140- ,f the loss sensitivity of this area is very high, exceeding the . capability of the plant. In addition, because of the weigh cell design, the material balan.cc can be very timely. Since much less Pu0 2 must be diverted than H0 2 , a f rmal material balance over this area appears both possible and highly desirable. - G 1 1 / ^ ~ - - - . . . . . ^ -= g +1 - ' ., l  ;~: Sa. ..mc f W' *' .a  ?, . 7b - '.' NLW ~~ ".- NT' ;a- -' ^ ' - ,%~ f - .i ; .=J. ~ - .;hh. g . ~ -k .; [b ,.Is,. .l .' ; ~ g_# ^ ;zw _; w. _ , r-141 REFERENCES

1. "Barnwell Nuclear Fuel Plant - Separations Facility Final Safety Analysis Report," Docket 50-332, Allied General ~

Nuclear Services, Barnwell, S. C. 29812, October 1973.

2. R. A. Schneider and D. P. Granquist. " Capability of a Typical Material Balance Accounting System for a Chemical Processing Plant," BNWL-1384, Battelle Northwest, Richland, Washington, May 1970.

3. F. M. Groth and F. O. Curton, " Evaluation of Instrumentation for Nuclear Fuels Reprocessinn Plant input Weicht Measure-ments." ICP-1014. Allied Chemical Corp., Idaho Falls.

Idaho, July 1972.

4. K. B. Stewart and L. F. Wirfs, " Analysis of Shipper-Receiver Data for Plutonium Nitrate and Plutonium 0xide "

BNWL-1682, Battelle-Northwest, Richland, Washincton, J a nu a ry 1973.

5. F. M. Smith, " Plutonium Dioxide Fced Materials Handlinn and Measurement Standards Studies," HEDL-TME-71-122, Westinnhouse-Hanford Engineering Develonrent Laboratory.

Richland, Washington, December 1971.

5. H. H. Ph. Moeken and H. Bokelund. "Some Developments in input Accountability at Eurochemic," l AEA-SH-133/77, Vol I, Safecuards Techniques, Proceedinos of a Svnposium, Karls ruhe, Germany, 6-10 July, 1970, international Atonic Energy Agency, Vienna, Austria.

7. " Evaluation of the Results of the PAFEX-1 Experinent."

Secretariat Working Paper, Co-ordina tion Meeting of Analytical Laboratories on the Review of Field Denon-stration in Process Analyses, International Atomic Energy Agency, Vienna, April 1974.

8. C. G. Hough, et al. , " Example of Verification and Acceptance of Operator Data - Low Enriched Uranium Fabrication "

l BNWL-1852, Battelle Northwest, Richland, Washington, i August, 1974.

9. 10 CFR 70.58.

10. "Barnwell Nuclear Fuel Plant -

Plutoniun Product Facility Preliminary San aty Analyses Report," Docket 50-332, , Allied General Nuclear Services. Barnwell, S. C. 29812, July 1974.

11. E. Bain, et al. , " Quantitative Evaluation of the Retimac System," SAI-75-579, Science Applications, La Jolla, Ca.,

June 15, 1975.

~

l l ~

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

[. ,, 7 P L -142-12 . Regulitory Guide 5.8. " Des i gn Considera tions f or Mi nimi z i n.! , Residual Holdup of Special Nuclear Materials in Dryino and Fluidized Bid Operations " U. S. Nuclear Regulatory Connission. Washington, D. C., May 1974.

13. Regulatory Guide 5.25: " Design Considerations for Minimizing Residual Holdup of Special Nuclear Material in Equipment for Wet Process Operations," U. S. Nuclear Regulatory Commission Washington, D. C., June 1974.

s 14 Regulatory Guide 5.26: " Selection of Material Balance Areas and item Control Areas." U. S. Nuclear Regulatory . Commission, rashington, D. C., June 1974. ,

15. Regulatory Guide 5.42: " Design Considerations for Hinimizing Residual Holdup of Special Nuclear Material in Equipment for Dry Process Operations," U. S. Nuclear Regulatory Commission January 1975.

16. J. L. McElroy, ed., " Quarterly Progress Report - P&D Activities, Weste Fixation Progr6m - April through June 1974," BNWL-1841 Battelle-Northwest, Richland, Washinoton, July 1974, p: 10.

17. H. A. Lewallen " Generic Hixed Oxide Fuel Fabrication Plant - Decommissioning Study " Draft BNWL Report, Battelle-Northwest, Richland, Washington, January 1975.

D f . ".w .. . ?W m-. ]-5{ - * - - i . . - _ . ..~ - c ~gj: ;p ==.. mxo m w ,. .{ - . _ . .. . .  :.g j; _ m e-I -143-APPENDIX C HTGR FUEL FABRICATION, PLANT MODEL The plant model for initial and reload core-fuel fabri-catioie is based on information provided by the General Atomic Corporation in support of their license application for the 6 proposed Youngsville, North Carolina, plant.III The Youngsville plant is designed to handle hinhly en-235 riched 0 feed and cannot fabricate recycle fuels. The plant is capable of fabricating six fuel cores / year. Initially, essentially all their production will be initial cores. In time, the requirement for replacement cores will increase. In this analyses it will be assumed that the facility will fab-ricate five initial cores and four one-f ourth core reloads in the year being analyzed. This condition nay be expected to exist in the late 1980's. The fuel management of these reactors is complicated by the requirement for no fuel shuffling. The standard fuel element shown in Figure 1, is a hexagonal block of graphite. In a 1170 MWe plant, these blocks are stacked eight high and are refueled seven stacks at a time.( } Heat generation rates 235 are controlled by varying the amount of U in a fuel assembly. Thus, in a core there cuuld be many enrichments. For this analysis, typical amounts of 235 U and thorium used in an initial core assembly and in a reload assembly will be used. The following sections describe the fuel assemblies and then the fuel assembly fabrication process. C.1 HTGR FUEL ELEMENT DESCRIPTIONS The standard fuel element for the commercial HTGR is shown in Figure C.I. It is a hexagonci block of graphite into which vertical coolant and fuel holes are drilled. The fuel holes are filled with rods composed of fuel particles (up to about . 60 volume percent) bonded together by a graphite matrix. A standard fuel element into which control rods can be inserted l is shown in Figure C.2. . .2 - - - -- - ~~ ^ - ~ ~ - - ~ ~ ^ s

• y

[~ / 7, b .. . *. , , ' .' YJ1 - " -. . J. $ . . . ~ ' mW -144-i A COOLANT H0LE ~ 21.0 mm DIA ' /$ ~ ~ ~ h s[9 [g._"g , .- o4 . a / JC BURNABLE POISON 359.1 mm h/ HOLE 12.7 mm DIA 23.0 mm Skjll  ? , ) ,' -FUEL HOLE PITCH ' 15.9 mm DIA (132) jy('M@Op9g/Idh_. ,'; CEMENT 4 \ c GRAPHITE FUEL HANDLING s . _9_.___.--. PICKUP HOLE PLUG (TYP) l- A n ,_ p DOWEL . / ' Nv @iY 'i:.li. ,! j! !):!  !.I j 8 HELIUM x .i l! !' s ,! FLOW ll i:ll

. 1 :ll!

-l 793.0 mm  !  ; .'d j ;} lll li - ,l'  !! : 'r : i s I<: i i l il l! :!

t. l

/' f i '  !}!!,! j!.5ili _ , ' , !: .e . : . .i SECTION A A DOWEL SOCKET . FIGURE C.1 HTGR Standard Fuel Element i -: 1 .- , = - - - -- ...f

f. J.' . su wy. . 4,p .

'OAW-L: $f ~ L.':5 :. '5,? .  ?.T] '9N*?* . l  ; L- ;:, ;l~G-.m - !. ,  : =-= -  :-- . -=.=- 43 .y. gP o i -145-  : P RESERVE SHUTDOWN HOLE 047 r- COOLANT HOLE .- 95.3 - 21.0 mm DIA mm 104.4 mm A I  : SOOSOO@ i + 101.6 359.2 mm 101.6 mm g,q 3 mm CONTROL ROD 23.0 mm ' CHANNEL PITCH V FUEL H0LE 15.9 mm O!A (76) PIA BURNABLE POISON V HOLE 12.7 mm DIA 00WEL PIN FUEL HANOM - 246.9 mm = PICKUP HDLE 4a, . o . 3,..,, , S __ i " 1 y; i: HELIUM  ! I, 3 FLOW w  ; (TYP) 793.0 mm ,1 , liii i if' - j

i l i!

,  : s I .- , j r h ) .. i A KET # SECTION A A INSTRUMENTATION CHANNEL  ; ~ FIGURE C.2 HTGR Control Fuel Element 1 s ...-. ..-- - .. . - -.... ~-. - . '_s p .,, ~~- ...~.. -w n.. r .. .-g...y. .t j- 'y-' -- - ~ ' ^ * ~ ' ~. . . ~.---e-... -s-. q yy_ ,_ ;; " +~ , , e- -146-The initial core is fueled with thorium oxide and 93% enriched 235 U as uranium dicarbide. Subsequent 235 reloads could either be enriched with 2 SU , recycled 0 233 232 or 0 bred from the Th. This analysis is not censidering 235 U recycle and 233 0 streams. the As currently envisioned the UC 2 is encapsulated in TRISO particles and the Th0 is encapsulated in BISO particles. 5 2 A microphotograph of the particles is shown in Figure 3. The TRISO particle consists of the inner fuel particle followed by successive layers of buffer carbon, pyrolytic carbon, silicon carbide and pyrolytic carbon. The BISO particle does not have a sic layer. Table C.1 shows the fue! particle size and the thicknesses of the various layers. The densities of the various TABLE C.1 HTGR TRISO and BISO Particle Dimensions Particle BISO TRISO Characteristic Particle Particle Fuel Th0 2 UC 2 Particle (microns) Dimensions Fuel Particle Diameter 500 200 Buffer Carbon Thickness 85 85 Inner Pyrolytic Carbon Thickness -- 25 sic Thickness -- 25 Outer Pyrolytic Carbon Thickness 75 25 , compounds making up the BISO and TRISO particles are shown in Table C.2. These density values are based on data presented  ; W 6 OM* me . . - _ _ _ , ,  % elke *[ - ]- ,- ---e* * ~-*T - 7mm - g, . ,5. 2 . $~.k ' d20'= _ $7 - I $. ' - Y.- -.$;5__ EN$ '$k h . . _ _ ' -- , ,.,.w.. , $,. . , ws.s a 0 I i I } t ! I t i t t'iM'y., ,% .s l . 4....4- ' .,4 } . 5 .1 - 1 *,'e "/k ,' h .- 2^ - OUTER ISOTROPIC d# ' - 4 .:ct 6 ;h . Y I7D . 40.\ / '/, PYROLYTIC CARBON '  ; ]h N w < . . n9 ' 5'"1 .  ?. (,.., . ; -y)s.'.' , '. \ { je e t 9 . , . , { ,  ; I.- SILICON CAR 8;CE * /, J.pdp . ,', */^

• b

SARRIER COATING ~. . .#  ; ~. '.,.. e . . . ..s e . e . r ~ ,  ? ..u.> .  !: e n. w: . I. g . . ,1 . . - ,- y.- ~ '. i Md *' ',","'",'5 - fr Ro< ,C1 %'*0N  ! W run a^ar'.ca  ; 8' \-1 T umman

• c j6fp3f!Ny < $

s  ; , ~ -.  ; .m p, , ;,44 ,, < *  % ,/ _ _ hs ' ' , . - < s ., /g , Y Olvf aC CARBON "" j ewn-

}

t , Nl er BISO TRISO h! j . g , FIGURE C.3 GA Coated Particles ' i '{t t, f .ai,it i $ s i if l I y . d t -  ; e 7 I -148-4 TABLE C.2 Density of Materials Used in BISO and TRISO Particles Material Dernity (iirn / rm3) , UC 11.446 7 A T h0., 9.H29 BufYer Carbon 1.1 sic 3.2 3 Pyrolytic Carbon 2.0 by P. R. Kasten et al'. I The fuel narticle density is based on a density which is 98" of the theoretical density. Based on these densities, the volume and mass fractioti occonied by the various materials are shown in Table C..i. TABLE C.3 BISO and TRISO Particle Volume and Weight Fraction Data BISO Particle TRISO Particle Volume Mass Volume Mass F ra c ti on Fraction Fraction Fraction Fuel Particle 0.2267 0.6448 0.0508 0.2345 Buffer Carbon 0.3188 0.0922 0.2709 0.1117 Inner Pyrolytic Carbon -- -- 0.1488 0.1227 Silicon Carbide -- -- 0.1888 0.2495 Outer Pyrolytic Carbon 0.4545 0.2630 0.3407 0.2816 Tctal Volome (Cm ) 3 2.887x10 -4 -- 8.245x10 -5 ,, Total Mass (gms) -- 9.978x10-4-- 1.997xio-4 The BISO and TRISO particles described in the previous paragraphs are mixed with a binder and graphite p' wder o and extruded into fuel sticks which are 2.48 inches long and 0.614 inches in diameter. Table C.4 describes tne aat:eun or a typical fuel stick, standard fuel assembly and control elener.t fuel , assenbly for both a total initial core loading and for a typical one-fourth core reload. This information forms the * ' T' *r- , - - - , . g- _W wn. - ij: 'iy .=eu . ;s. . .J.= &. _ ,_ . R. _X . . . _ . _ 7 P -149-TALLE C.4 Description of Core Compositions for a 1160 MUc itTGR Component Initial Core Reload Core Compusition Description Composition $-1/4 Core Reloads) Fuel Sticks Length - in 2.48 2.48 Diameter - in 0.614 0.614 Th Content - gms 6.414 5.774 Th07 Content - gms 7.299 6.570 2330 Content - gms 0.277 0.499 UC C ' 0.330 0.593 Vo$"ontent-ams BISO 27.2 24.5 Vol ". TRISO 4.8 6.7 Vol % Natrix 68.0 66.8 Total Wt - gms 21.73 21.560 Std. Fuel Assembly Number / Core 3360 3360 Fuel Sticks / Assembly 1581 1581 Th02 - kgs 11.540- 10.387 UC2 - kgs 0.521 0.938 sic - kgs 0.542 0.988 Burnable Poison - kgs 0.907 0.907 Carbon Coatings - kgs 8.005 7.790 Fuel Stick Binder - kgs 14.230 13.980 Graphite Block - kgs 83.550 83.550 Total wt - kgs 119.295 118.540 Controlled Fuel Assembly Number / Core 548 548 No. of Fuel Sticks / Assembly 909(681) a 909(681) Th0 2 kgs 6.644 5.980 UC 7 - kgs 0.300 0.541 [ sic - kgs 0.312 0.568 l Burnable Poison - Kos 0.605 0.605 l 4.609 4.485 ! Carbon Coating - kgs , l Fuel Stick Binder - kgs 8.137 8.037 l Graphite Block - kgs 81.246 81.246 Total wt - kgs 101.896 101.462 Total Core Thorium (kgs) 37,486 33,739 2850 (kgs) 1.617 2.913 l a) Bottom Fuel Assembly in Control Rod Stack (73 total) l ~ i _ i l - - - - - - - - - - ~ m-g .. .- k w ws ,. Y ,., ~,7 .: ,. ..z. . ~ ~ 3[g, _=. SY Y aw " -150-basis for the material accounting evaluation of the fuel f abri-cation plant shown in the following section. The next secticn will present an abbreviated descritpion of - a material balance taken over the fabrication plant. This will be followed by sections on measurement uncertainties and firally a section summarizing the results of .the material accounting calculations. C.2 MATERIAL BALANCE MODEL FOR AN HTGR FUEL FABRICATION PLANT The fabrication of HTGR fuel is basically a batch process. Criticality requirements limit inventories at most locations 235 Batch processing is to a few kilograms of contained U. one of the most straightforward ways of controlling inventories in the facility. In most processing areas, the batch size is 235 , limited to 3.6 kg of 0. Most conversion steps require that a batch be transferred out before another can be transferred in. The abbreviated description of the plant flows is shown in Figure C.4. Based on the data provided in the previous section and the Youngsv111e license application information, the material balance is summarized in Table C.S. The scrap generated during the inventory period is assumed to be held over into the next period. This scrap includes furnace liners which are one of the major sources of measurement uncertainty. t f -m-. m - om . i

e. . .. .. . . . . g,_ , . ge .. .jg a ,* ,,

g * - N' & ~~ ' y ,f{*  ;-h .

,+ {.; ,y f,  != & =  : '- Ti=, 9E +  :==; =ws; ~yy - -~-: 7 P -151-TABLE C.5 HTGR Material Balar.ce - 2 -Month Accounting Period

• Initial Core Campaign Batch Size No. of Total Batches kg F,1,ow Conoonent kJ 23*U UF6 to Converter 15.55 105.185 1632.465 UO F 3.6 453.35 1632.074 i, 22 Fresh UO2 Feed 3.6 452.51 1629.028 Recovered U02 (from scrap) 3.6 133.41 480.289 Mix Line 0.72 2929.C 2109.317 UC2 Consert 0.72 2844.4 2047.973 UC Sperodize 3.6 552.4 1988.529 2

Buffer Seal Inner LT1 Coat 3.6 538.4 1938.748 ~~ Sic Coat 3.6 520.69 1874.489 Outer LT1 Coat 3.6 505.55 1819.948 Blend 3.6 475.69 1712.492 Green Rod 0.277 qms 6.182,282 1712.492 Carbonize 0.277 gms 6.007,282 1664.017 Final Product 0.277 9ms 5.832.282 1615.542 Element Product 0.4092 3944 1613.905 l G M ~ h eew . _ , ., e-_- - 4dhmeMethseemthhe e me.Eue r .e ,- %, * ' 6 b ew" 4 p as , 1 79 P" l i i -152-C.3 ESTIMATED MEASUREMENT UNCERTAINTIES FOR AN HTGR 1 FUEL FABRICATION ?Lt.NT 1 For the plutonium processing facilities, there was a wealth of data from independent sources which were used to obtain estimated measurement uncertainties. Because of the unique I characteristics of the HTGR fuel, there is much less relevant data. Table C.6 presents a summary of the estimated capability e of HTGR neasurement sys tems. The categories CU-l through CR-14 designate streams shown in Figure C.4 This information was obtained by personnel communication with GGA personnel. 1 The designation short-term and long-tern systematic error needs some clarification. Although the definitions correspond to the usage described in Appendix A, the short-term error is based on weekly recalibrations. Thus, in order to get the systematic error component associated with a 2-month period, the methuds developed in Appendix A must be used. For the evaluation o' the capability of an HTGR fuel processor to account for the processed 235 U in a 2-month accounting period, the short-term systematic error component will be combined with the random error term. In addition to the well measured, in-process i nven to ri es , holdue and measurement uncertainties associated with scrap recovery must also be considered. ' Table C.7 presents an evaluation of the estinated levels of holdup in equipment and these estimated u measurenent accuracies. At the present time the processes - to be used for scrap recovery are incompletely specified. As l a r e 'o u l t , the data presented in Table C.7 are considered to be I l l m ._ - - - _ _ [ I1 -

= is

. v~  ; et .- - -pe- _,-w.R. zQ ;._ . m :- .w m~ .xQ:p :y- ~ Y..W : - .z..=.. q.;.p f.e -: X " t.. rW3 L 2, l -153-TABLE C.6 HTGR Fuel Error Components Shott Long Random Term Term Error Syst. Syst. ~j Category _ Methed Cravimetric .01 - .03 U038 - UO2 .01 . Mass Spec .02 - XRF 2.25 . 0.4 0.3 ITIh0xide .01 Mass Spec .02 - XRF 2.75 0.4 0.3 CU-1 XRF 2.6 0.4 0.3 CU-2 - 0.3 C"-3 (Coop) XIF 1.0 0.4 XKF 4.0 0.4 0.3 CU-4 XRF 2.6 0.5 1.03 l CU-5 XRF 3.0 0.5 1.03 CU-6 CU-7 (Comp) XRF 1.0 0.5 1.03 2.0 0.5 1.03 CR 13 XRF 1.04 CK-14 Delayed Neutron 2.0 0.5 Activation Analysis Scrap-Liquid KkF 2.75 0.4 0.3 Scrap-Solid Cam a Count 4.0 1.0 2.8 Waste Barrels Carmaa Count _ 6.0 2.8 24. Duct Holdup Casmaa Count 5.0 NA '20. . Filters Cam.a Count 6.0 NA . 53. l l _m _ 1 1f j - .r e - 1 - cir - .. . ; ~ _. .- _, , _ , , . . , <] I -154-us e ii UOgig WASTE UO3 t t I tus tint SCAAP eWA$ff

• MECOVERY .

<i CONVERT g CU-2 o - l sewnoiDirE _ CU-3% auff(R.54AL. IN8sER til COAII8eG C_U- 4 $ se COAtems0 CU-5 i OUT ER LTI COAftNG CU-6 i etENo etniita PAAllCit$ A CU-7 V GREEN ROD { CR-10 cAnsoNin b e CR-11 HCI Ctf AN I CR-13 SAKE o CR-14 LOA 3 ELEMENT O it i FIGURE C.4 Fissile Production Sequence Sumary L _ -,.- . . . . .v -- ~ ,wi ti~~~ * *~..:. ~ Q '2.- n -, .a T'. ~* 2*3~ hcMi ,!bhmy' h .1[Uh '  ; -.~;t[= c -[, . $*f_2 - ~,y'y' _T * .* ; cpg. y gy, ,gg ~, .i.Q, ~ .;?g. . - - . < ;~~- ;- .-c...-.-< - . ,_7 7F -155-TABLE C.7 Inventory Levels and Estimated Measurement Errors Estimate Number of Inventory Con ponents '"U Per Measurement Accuracy (or Batches) Component Random Systematic Component 7, Y, (oms)  ? In-process Equipment 4 3600 10 -- U02 Conversion Units 6 360 50 -- UO -C Blending 2 50 -- 18 36 ,UC Conversion 2 50 -- UC Spheredizing 9 300 2 50 -- 3 250 Screens 36 20 50 -- Coaters 50 <1 50 -- Stick Fabrication Failed Equipment 400 20 20 -- Coater Liners 3 300 20 -- Conversion Units 100 100 20 -- Hisc. Equipment Scrap fast Scrap 45 3600 2.8 -- 88 3600 4.0 -- Off Speck Particles 15 409 2.3 -- Defective Elements 500,000 0.277 2.3 -- Off Speck Fuel Sticks 500,000 0.277 2.3 -- Green fuel Sticks in-Process Materials 60 3600 0.02 0.01 U0 l 2 3.0 0.3 50 720 ll UO2 -C Blends 60 3600 2.8 0.3 UC2 Powder 50 , 3000 1.3 0.3 l UC 2 Particles 10 3600 4.0 0.3 LT1 Coat 20 3600 2.9 1.03 sic Coat 60 3600 3.3 1.03 5 Triso Inspect 28,800 0.277 1.7 1.03 Triso Blend 86,400 0.277 2.4 1.04 Stick Fab L-f - g j 1 b ^ * . .. $; ~ .' ..~. 5r -156- - order of magnitude estimates only; they are probably conserva- , tive. These estimates were based on studies of plutonium holdup and also from available descriptions of HTGR processing facilities. 235 U HTGR FUELS PLANT C.4 MATERIAL ACCOUNTING CAPABILITY OF A The previous sections have provided all the information a required to evaluate the capability of material accountino 235 U-HTGR fuels plant. The variance cal-in a high enriched culation for the inventory term at the beginning or end of a 2-month period is shown in Table C.8. There may be material that was not processed during the 2-month period, such material was not included in the variance calculation. The inventory variance is included with the flow variance term in Table C.9 to obtain o(MUF) and LEMUF estimates. These estimates were used to provide a basis for comparison of HTGR f acilities with plutonium processing f acilities. This comparison is shown in Table 4.1. The base case. 2-month accounting period was studied. , Improvements in sensitivity and timeliness were beyond the scope of this study. Such an evaluation would require a much more detailed description of the scrap recovery operation, t 9 I i 1 i J _ .m_ _ t_ _ _. . '??*' .-llr '. * * * # W- ---r,--- - ..;~,. .~. . , ._ _ **g 8 .M E,, '; * * .T . ' .,4 ""**.L.  ; Q fLj * :::id;\ . = ;5 . , ,2?-h , . r,~Q n_ w ._. 7e l 157-TABLE C.8 Summary of Inventory Uncertainty Calculations 235 U Content Varia ce Inventory Component kg's kg In-Process Inventories y UO Prep 212 0.034 2 UC2 Prep 432 1.427 Triso Particles 324 5.407 Fuel Sticks 32 0.029 Total in-Process 1000 6.897 Scrap Recovery Inventories Failed Equipment 20 6.594 Coated Particle Scrap 320 1.825 Uncoated Particle Scrap 160 0.457 500 8.876 Equipment Holdup 20 0.968 TOTAL INVENTORY 1520 16,741 e e 4 - m .m  % b

• e- ,

w a *'g 3 + D.. Y y -158-35 g TABLE C.9 Summary of Variance Calculation for an HTG'l Fuel Fabrication Plant for a 2 -Month Accounting Period Component Variance Component __ ka2 235 g Feed - 15.52 kg's of No. of Satches 105.184 p Random Error Variance 0.073 Systematic Error Variance 8.634 i Total Variance - Feed 8.708 [ Product - 0.277x10~3 kg's No. of Batches - 583.2282 Random Error Variance 0.0003 Systematic Error Variance 276.892 Total Variance - Product 276.893 Wa'ste - 7.4x10-3 kg's No of Batches - 2500 Random Error Variance 0.0009 Systematic Error Variance 19.842 Total Variance - Waste 19.843 Total Flow Variance 305.4 Inventory - 1520 kg's Inventory Variance (Beg. & End) 33.5 2 338.9 Total Variance - kg sd(MUF) - kg's 18.4 LEMUF - kg's 36.8 LEMUFx100/ Feed 2.25 LEMUFx100/(Feed +1ny) 0.86 k i i . \ i I i f . I . . . . _.-.--_..m... _ . . . . . ' ~ ~ ' e sp a ;p,;; 5;g sy;g 33 ifgy; .a_g;. . J::e f I l i F -159-  ! REFERENCES FOR APPENDIX C (1) Environmental Impact Report - Gulf Youngsville Nuclear Facility, DCCKET NO. 70-1372, February 1973. (2) R. C. Osh1 berg, R. F. Turner, W. V. Goeddel, "HTGR Fuel and Fuel Cycle Summary Description " GA-A12801 (Rev), General Atomic Company, San Diego, California, January 1974. I P. R. Kaster, et al. , " Gas Cooled Reactor Programs," (3) ORNL-4911. Oak Ridge National Laboratory, Oak Ridge, Tennessee, 1972. F (4) H. E. Stelling, "23:Refabrication 0 Particles,"Costs for HTGR GAMP-9478 GulfFuels Containing General Thorium-Diluted j Atomic, San Diego, California, September 1969. (5) H. E. Stelling, et al., " Central Refabrication Plant for HTGR Fuels," GAMP-9145, Gulf General Atomic, San Diego, California. Harch 1969. (6) HTGR Base Program - Quarterly Progress Report for Period l Ending February 28, 1974, General Atomic Company, GA-A12916,  ; San Diego, California, March, 1974. i 1 } 4 l l ^^ Y 'a p $ q ,% k,:h-hkk.h. ***m' =.w g.=.= .: = g. .,_-_g # t i *,hA ft,s o "ig P EH j Q I; @4 hj L " El tf ~. 14 e I9 Mf. l.4' 1 .l Lf1 h,1 [4 y4 rM [+ A N , F1 a t .i l pj t.. .. < a ,5) ,:p -;a 4 " Jg t b;.s4! 3 a 1 m  :. #gT2.e . 4:.j$,

u. y f

fii L j . .i .sj c4 E 1 L DOM $-S j -m ,? - in ' g - G I,* t .S h esdh4466 p'.. $@W JT MM 'NNN - N -We e

• 4 A* = '
• U # - - ' - -M. e = a

, e e .g ,,,,. g- , s-e . ' , e' ) ~' o., ,- n. -m .w. ; .. = a, ii3nE r.i.c-si!55555sWEH M3s "- 4 .- : - - ~ . - ~ -dw d .t  : uw -M5 .:i :E .tg c_,r.=:-:, . . mn =.e rwa,m.- _, _-s-,. < ::.:~r m..-_.e $5 .' .:) 7,1., _5iEI_ . , ,7-j;c-- _ --._ ws; .; Wh - -,}}