NUREG/CR-6410, USEC - Staff Pre-filed Exhibit 25, NUREG/CR-6410, Nuclear Fuel Cycle Facility Accident Analysis Handbook

USEC - Staff Pre-filed Exhibit 25, NUREG/CR-6410, Nuclear Fuel Cycle Facility Accident Analysis Handbook

ML070800427
Person / Time
Site:	07007004
Issue date:	03/31/1998
From:	NRC/NMSS/FCSS, Science Applications International Corp (SAIC)
To:
References
ASLBP 05-838-01-ML, Job Code J5074 NUREG/CR-6410
Download: ML070800427 (23)

Text

NRC STAFF EXHIBIT 25 4 42 NUREG/CR-6410 Nuclear Fuel Cycle Facility Accident Analysis Handbook Manuscript Completed: March 1998 Date Published: March 1998 Science Applications International Corporation 11251 Roger Bacon Drive Reston, VA 20190 Prepared for Division of Fuel Cycle Safety and Safeguards Office of Nuclear Material Safety and Safeguards U.S. Nuclear Regulatory Commission Washington, DC 20555-0001 NRC Job Code J5074

• 10

A 3.4 Source Term Estimations for Inadvertent Nuclear Criticality Excursions A nuclear criticality accident is defined as the release of energy as a result of inadvertently producing a self-sustaining or divergent neutron chain reaction. In fuel cycle facilities, a criticality accident may occur following: (1) transfer or leakage of fissile material from a geometrically favorable container to a container with unfavorable geometry; (2) introduction of excess fissile material into a container or a. ca; (3) over-concentra tion of a solution; (4) failure to maintain adequate quantities of neutron absorbing materials; (5) precipitation of fissile solids; (6) introduction of neutron moderators into a system or area, and (7) alteration of system or area geometry.

A nuclear criticality differs from othe~r fuel cycle facility accidents in that radioactive material is generated in the event and short-lived fission products not normally encountered are present. Nuclear criticality events may be terminated by control systems or by thermal expansion, loss of fissile material or moderator mass, change in density, or mixing caused by the rapid release of energy.

Criticality accident hazards include direct exposure to prompt neutron and gamma radiation produced in the criticality event, release of gaseous radioactive material, and expulsion or entrainment of radioactive aerosols. Direct exposures in the immediate vicinity of the event can be substantial, but off-site impacts are generally below detectable levels.

For nuclear criticality safety (NCS) evaluations, systems are categorized as solutions; moderated/reflected solids; bare, dry solids; and large storage arrays. The total amount of energy released in the event is generally small, and physical damage to ventilation and gas clean-up systems is not expected to occur.

Mixing within the confining building and normal function of off-gas systems delays and mitigates the consequences of the event. The time during which a criticality occurs and the severity of the event depend in a complicated manner on the quantities, physical and chemical form, and concentrations of the fissile material and on the size, configuration, moderation, reflection, and neutron absorption characteristics of the system.

NRC requirements for NCS at fuel cycle facilities include adherence to the double-contingency principle in process design and operation, thus reducing the probability of an accident to a small value. Because of application of the double contingency principle, criticality accident scenarios involve multiple equipment or control system failures and are of such complexity that generally applicable analytical tools are not available. Thus, the nature and magnitude of possible accidents are assessed individually and conservative analyses are used to evaluate the adequacy of NCS protection systems. The balance of this section discusses the magnitude of possible criticality events at fuel cycle facilities and presents a representative method of source term estimation using a solution criticality at a uranium fuel fabrication facility as an example case. Other aspects of accident analysis for nuclear criticality events are discussed in a sample problem presented in Section 9 of Appendix D.

3.4.1 Fission Yields Thbe amounts of energy and radionuclides produced in nuclear criticality accidents are proportional to the total number of fissions occurring in the event. Thus, estimation of the total number of fissions provides a necessary basis for estimation of accident source terms. Estimates of the total number of fissions have been based on review of criticality accidents (Stratton 1989 and Frolov, et a] 1995), reactor excursion experiments (Nyer, et al 1965), and other experiments. Reviews of these data and empirical models developed primarily for solution systems (Olson, et al. 1974) provide perspective on estimates of total 3-93 3-93 NUREc3/CR-6410

number of fissions and energy generation rates. Descriptive information and data reported in the literature for a variety of systems are summarized in Tables 3-8, 3-9, 3-10, 3-11, 3-12, and 3-13. These data and estimates are discussed below for the four major types of reactive systems. Mathematical models capable of estimation of criticality accident progression and termination require coupled representation of hydrodynamic and neutronic phenomena. Although such models have been developed for power reactor system, models capable of estimating energy generation and total number of fissions have not been developed for the diverse systems encountered at fuel cycle facilities.

3.4.1.1 Solution Systems Solutions are of principal concern for NCS evaluations at fuel cycle facilities because liquids are mobile and can be formed into geometrically unfavorable shapes with inadvertent moderation. Data for accidents that occurred in the U.S. are presented in Tables 3-8 and 3-9. Data for accidents that occurred in the former Soviet Union are presented in Table 3-13. Of the 13 accidents listed in Tables 3-8 and 3-9 that involve solutions, 8 occurred in process facilities whereas 5 occurred in nuclear criticality experimental facilities. Inadvertent transfers to non-favorable geometry vessels and failure to follow or properly interpret procedures were the most common causes of the accidents. The total number of fissions for the accidents ranged from lxlO'5 to 4x1019 and an initial burst was generally followed by a plateau period characterized by a lesser and declining fission rate. Because of the duration of the plateau period, the major portion of energ release occurred during this time. The highest fission yield (4x10' 9) occurred when a relatively large volume containing approximately 70 times the critical mass (i.e., 34.8 kg of U-235) was transferred to a vessel of unfavorable geometry positioned above concrete, which reflects neutrons. Of the 12 accidents listed in Table 3-13, 11 involve aqueous or organic solutions. The accidents generally had an initial burst followed by repeated smaller bursts extending over periods of time of the order of hours.

Of these events, the largest had a yield of 7.9x10" fissions and involved approximately 2.8 kg (6 lb) of highly enriched uranium.

3.4.1.2 Fully Moderated and Reflected Solids Estimates of peak fission rate and total number of fissions for an accidental nuclear criticality in a moderated, reflected solid system may be derived from data from accidents and from experiments with light-and heavy-water-moderated reactors. Criticality accident data are reported in Tables 3-8(b) and 3-11 for uranium and plutonium elements of various shapes with water or graphite moderation. The reactor excursion data reported in Table 3-10 is for uranium-aluminum and UO2 stainless steel clad fuels. The total number of fissions for the relevant accidental criticalities reported in Table 3-8(a) ranges from I x I015 to Ix 10 while the total number of fissions for reactor excursions is bounded by 5xl0'" for the power levels reported in Table 3-10. Criticality events in moderated, reflected solid systems were characterized by an initial burst with little or no plateau period.

3.4.1.3 Powder Systems Limited experimental data exist for powder systems of the type found in fuel cycle facilities. Data from accidents presented in Tables 3-8(c), 3-11, and 3-12, for systems that may serve as surrogates for powder system, are bounded by a total number of fissions that is less than 2x10'7.

NUREG/CR-64 10 3-94

Table 3-8. Summary of Known Accidental Criticality Excursions (1945 to 1974)

(a) Solution Systems, (b) Metal Systems, and (c) Moderated Foil and Powder Systems (Olsen, et al. 1974)

No.

Data Locationl Tii l b Jaera Arrnjmeoni Duratio Towa J Cusepbyi (a) Solution Systems SE 1 12149 LASI, NM Water U(93)O2(NO),

Sphere.

- 3x10RI Not 3-4x10" Control rods None Boller

(- I kg U235; 13.6.1) graphite-(barely c0,'r known withdrawn reflected prompt too fast critical)

SE2 11/16/51 Hanford works -

P-Il PuO2(NO,),

Spbes. 93 percem 8x10 Single Rx10" Too high None Richland, WA (1.15 kg Pu; 63.8 I) fuifl-reflecied burst fuel addition SE 3 5/26/54 ORNL, TN Spicdr UO1F3 (18.3 kg U235; Cylindrical 5xI0" Not IxlO" Shift of None 55.4 i)

annulu, known poison umreflected SE 4 211/56 ORNL. TN Scram UOF, (27.7 kg U235;

Cylinder, 1.6x10" Single 1.6x10" Geometry Warping of blade 58.9 I) unreflected burst change bottom of

_+

cylinder SE 5

/30/68 OILN. TN U-233 UO,(NO),

Spbe, water-L.a0l10 Single

,lxiO" Air in line None

- 1 kg U233;5.8 1) reflected burst P 1 6/16158 ORNL. TN -

Y-12 UO2(NO,),

Cylin,.er.

- L.1x10" 13 min.

1.3x10I Valve leaked None Y-12 Processing (2.5 kg U235; 56 I) concrete-or left open (os: $1000)

Plant reflected below P 2 12/30/58 LASL. NM - Pu Agitator PuO,(NO,),

Cylinder.

l.SxlO" Single lx10" Procedure not None Processing Plant (3.27 kg Pu; 168 1) water-burst followed reflected below P 3 I10/16(59 Idaho Reactor IF-I UO,(NO),

Cylinder, 101 Not

-4x10`

Sparge gauge None Testing Ajrea, (siphon)

(34.5 kg U235; - 800 i) concrete-known plugged (loss: $62,000)

Chemical reflected below Processing Plant LA 0b

C)

Table 3-8. Summary of Known Accidental Criticality Excursions (1945 to 1974)

(a) Solution Systems, (b) Metal Systems, and (c) Moderated Foil and Powder Systems (Olsen, et al. 1974) (Continued)

-OAM Fkk T

P 4 1/25/61 Idaho Reacor IF-I1 UO,(NO,)

Cylinder No estimate Not 6x107 Instruction None Testing Arm (air lift)

(8 kg U235; 40 I) known misintrprete (koss: $6000)

Chemical d

~~processing~

Plant_

P 5 417/62 Hanford works -

Recuplex Pu complex (15 kg Pu)

Cylinder,

- 101' 37 hbr 8Xl0" Valve leak None Richiand, WA tuaflected or opened (loss: $1000)

P 6 7/24/64 Wood River Wood U10(NO,),

Cylinder, I.ixl01" Not 1.3x10" Procedure None JunctionRI.

River (2.64 kg U235) unreflected known not followed sco~p recovery P7 8/2470 Widscale Works, Windscale Pu compkx (- 2.5 kg Cylinder No estimate 5-10 sec lx 10" Pu None Fngld Pu; - 100 t) accumulated

_in organic (b) Metal Systemrs ME 1 6/4/45 LASL NM Metal 83% U235 enriched U Array of cubes;

- 3x10" Not

-3xI0" Woaf leaked None Cubes metal A in. cubes waler-reflected known into array (35.4 kg)

(perhaps 3 bursts)

ME2 9/21/45 LASL NM Dragon Delta phase Pu metal Splere reflected

- I.Sx10"

< I sec

- IxlO" Dropped Nowe (6.2 kg) by Be (10 cents over) reflector block ME 3 5/21/46 LAS4 NM Screw-Delta phase Pu metal Sphere reflected 1.8xlO" Not U3x10" Screwdriver None driver by WC known holding reflector -way fron Pu slipped ME 4 2/I/51 LASL, NM Aquarium Two cylinders U(93)

Side by side in

-6xl0" No:

IxtO" Went critical Slight machine metal (24.4 kg and 38-5 water tank known during oxidation kg)

(perhaps practice several scram bu-sts)

Table 3-8. Summary of Known Accidental Criticality Excursions (1945 to 1974)

(a) Solution Systems, (b) Metal Systems, and (c) Moderated Foil and Powder Systems (Olsen, et al. 1974) (Continued) 7ý:I Wer 7-Dote I

L41cagid.~,

ebc RE U-111a I 1~Uq 4r~

jp-1,0 ME 5 4118/52 LASL, NM JaUima U(93) metal (92.4 kg)

Cylinder,

- lxIO"

< I sec 1.5xl0i' Computaon None unrclfled (21 cents ova')

error ME 6 2/3/54 LASL NM Godiva I U(93) metal (53 kg)

Spbere.

5.6xl0" Single 5.6x 10" Assembled Slight unreflected burg:

too rapidly warping of (loss: $600)

ME 7 2/12/57 LASL, NM Godiva II U(93) metal (54 kg)

Sphere, 1.2x10" Single L.2z10" Graphite fell Warping unreflected (21 cents over) burst against oxidation

=-ar meing close to -

e

(_oss: $2100)

MES 6/17/60 LASL NM 9-inch U(93) metal (48 kg)

Cylinder.

- I

'IO" Not 6x404 Error in Trivial cylinder graphite-known addition

______efieceedsima ME 9 1/I/W61 ORNL, TN U-U(93) metal (-'75 kg)

Cylinder.

- lxI0" Not

- lxlU "

Error in None Parffin paaffin-known addition reflected estimate ME 10 3/26/63 LARL, CA LRL U(93) metal (47 kg)

Cylinder,

]xIO 7 Not 3.8x1o" Ram caught Metal melted Be-reflected known reflector.

and some fifted; fal burned; contam~ination (loss: $95,000)

ME I1 5/28/65 WSMRI NM U-Mo U(93)-1.0% Mo (96 kg)

Cylinder.

ISsl0" Not 1.5x101 Incorrect Assembly bolts Alloy unrefleczed known operation broken. minor damae _o coating 0) 0

C C)

Table 3-8. Summary of Known Accidental Criticality Excursions (1945 to 1974)

(a) Solution Systems, (b) Metal Systems, and (c) Moderated Foil and Powder Systems (Olsen, et al. 1974) (Continued)

_firs An ml I

Lortio 16c on~~aua Duain Toa U

(c) Modeated Foil and Powder Systems H 1 2)11/45 LASL NM UHM -

U(93)H, pressed in Assembly of

-6x10I Single

-6x10" Excess Cubes swollen Styrex Stynf (UC.H,)

blocks burst mcdvi-y and blistered blocks addition H 2 713/56 IASL, NM Honey.

U(93) metal foils Cylinder.

Not Not 3.2x10"'

Assembled None comb sandwiched with Be-mflected known known too carbon rapidly H 3 12/11/62 LASL, NM ZEPO U(93) meal foils Cylinder. C-3x10" Single 3x10" Ewess fuel None sandwiched with and Be-(12 cents burst addition carbon reflected promp critical)

Table 3-9. Accidents In Proces ants (Paxton 1980) 6/16/58 Y-12 I.3x101s

- 7xlO16 365,339, "U solution

327, washed into drum 270, 236, 69, 69, 23 12/30/58 LASL 1.5xIO" 1.5x10' 7

- 4400 Pu concentrated (fatal),

in solvent layer 135, 35 10/16/59 Idaho CPP 4x10" W

107t 50 R, 32 R, "U solution mostly beta siphoned into tank 1/2/61 Idaho CPP 6x1017 6xl007 None 25 U solution forced into cylinder by air 417/62 RECUPLE 8.2xI0I' 1016 87, 33, 16 Pu solution in X

sump sucked into tank 7/24/64 Wood River 1.3x1OI

~ 10-10,000 (fatal),

235U solution Junction two 60-100 poured into tank 8/24/70 Windscale 1015

- 10i5 Negligible Pu concentrated in trapped solvent 10/17/78 Idaho CPP 3x101 Unknown None

`5 U buildup in diluted scrub solution 3-99 NUREG/CR-64 10

C),

F Table 3-10. Destructive Power Excursion Summary (Nyer, B!i~btand McWhorter 1965)

VASdt UTivty"

, e".

Pek EmmW4 '~

~m a

BORAX I 28]

3.1 384 s 19,000 135

<1.8m0 s 6,500 6,000-10,000 Destred cam. vasel, mad som associated eqpi-entL Stmall fission-product release. Stearn explosion

___proposed ascam.

SL-I [291 3.0 280

- 19,000 133

>2,075

>7,300 10,000 Destroyed cme, bulged vesse, local fission-prduct

-i 10% flssaooio-ipod

• eiem1 Steam explosion - mi contribution from mtal-HO reaction.

SPERT-I (30]

2.6 200 1,130 11 585 2.000 7

Meed > 0.5% of cor.

D 12M25 2-7 218 1,270 19 680 2,300

.8 Melted - 3% of coe.

3.55 313 2,250 31 i,360 4,600 s4,000 Meked - 35% of coxe Desuavd cam and assommied equIpkmee bulged ftL

- 4% fli5-pnuldct meea PmbbC stemn expk*tm

- AO, malys bdicntes - 3.5 MWs amgy reease from.eton0 rtaon.

SPERT 1[3 11 2-6 455 17,400 155 1.800 2,200 70 oxide core 130 Two fuel rods nwupe Dmkrmdmo and/* r 3.3 645 35,000 155 1,800 2.200 defdcmaic of 2.5% of fuel rods. Negligible fission-produc releme SNAPTRAN-3 3.5 1,400

-20,000 50

>2,500 7,100

-4.000 Burs prsnur vesse..Al fuA W roL nrptured. - half of

[32]

luel reduced to powde form. Negligible fission-product release.

0 Table 3-11. Inhomogenous Water-Moderated Systems (Stratton 1967) 6/4/45 Los Alamos. New Mexico 35.4 kg U Pseudo-phere, water-

-3 x 10" Water seeping None

- 93% L"U reflected between blocks

% in cubes 2/1/51 LASI. New Mexico 2 cylindes U 2 cylinders. water-10"7 Scram increased Slight 24.4 and 38.5 kg 93%

reflected reactivity otidaion 7/6/52 ANL. Illinois 6.8 kg cunl oxide labomogeneous I.22 x IV' Manual withdrawal Plastic particles in plastic cylider water-of central safety desroyed reflected rod 12/12/52 Crlhk River, Canada Normal U Robs, DGO-modenaed, 1.2 10 Safety circuits core piteiflcted failed; control rod rined 7/22/54 Reactor Testing Area, U-Al plates, Al clad Inhomogencous 4.68x I0"'

Estimate of Reactor Idaho Fall, Idaho cylinder. water-expected excmion destroye reflected too low 10/15/58 Vinca, Yugoslavia 3996 kg Rods 13O0-moderated.

2.5 x 10" Too mch DO None Normal U uneflected added ian final step.

ofexperiment 3/15/60 Suclay, France 2.2 tons UO, Canned UO, rods in 3x 10" Control rod None 1.5% eniched water withdrawn 111/61 Reactor Testing Area, U-Al plate. Al clad Inbornogeneous 4.4 x 10" Quick manual Reactor Idaho Fall&, Idaho cylind, water-withdrawal of destroyed.

mdrtdcontrol rod building I

contaminated 1213*65 Mol. Belgium 1.2 x i10 g UO2 Canned UO, rods in 4.3 x 10" Manual withdrawal Nore 7% enriched HO-D0 OII of control rod 0

q0

)

• a 0

0,,

Table 3-12. Miscellaneous Systems

~~(Stratt l 97)____

2/11/45 Lx Alamos, New Mexico UM, pesed in styret Cylinder

-6x 10" Reflector added UH,-styrx cibes and/or source too swollen and lrge blistered 1953 USSR Pu(NO,),

Block tank 2.5 x 10" Tranfer to unsafe None Idaho Reactor Testing Area iin "U rods Cylider, rods 4.7x 10" Inwrect suam used Cow moften NaK cooled 7P3/6 LASL, New Mexico 58 k U Cylinder 3.2x I0Y' Change onk from None 93% m U, 2-and ICiOtbly 5-mil foils too la 11118156 Reactor Teasing Area, 1'"U Ni-Cr eements, Cylinder, protctype 2.51 10"'

Incorrect wiring in Every fuel Idaho Falls, Idaho ZrH-moderaed aircauft engine im chamber circuit cartridge melted 12111/62 LASI. New Mexico 2'U foils in graphite Cylinder. graphic 3x 10" inadequate None and Be-reflected communication between work crws

Table 3-13. Nuclear Criticality Accidents at Russian Industrial Facilities (Frolov, et al. 1995)

Date Plntf ThtalIs.1on i

Not-s 3/15/53 Mayak Enterprise 2.5xiO'7 Transfer of Pu solution into unsafe geometry vessel.

4/21/57 Mayak Enterprise 2x1017 Unmonitored accumulation of uranium oxolate precipitate.

1/2/58 Mayak Enterprise 2.3x 1017 Workers tip vessel, creating unsafe solution geometry.

12/5/60 Mayak Enterprise 1x10 17 Faulty plutonium mass analysis, accumulation of solution and precipitate in unsafe geometry vessel.

8/14/61 Siberian Chemical IxI106 Unmonitored accumulation of uranium Combine hexafluoride in oil vessel.

9q7I62 Mayak Enterprise 2x1017 Unmonitored addition of Pu scrap to dissolver vessel.

1/30/63 Siberian Chemical 7.9x 1017 Error in measuring uranium solution Combine concentration, transfer to unsafe geometry vessel.

12113/65 Siberian Chemical 2x1017 Unmonitored accumulation of uranium Combine solution in unsafe geometry vessel.

11/13/65 Electrostal Fuel lxlIOs Unmonitored accumulation of U0 2 Fabrication Plant slurry in holding vessel.

12/16/65 Mayak Enterprise 7x1017 Faulty accounting, loading of uranium solid into a dissolver with unsafe geometry.

12/10/68 Mayak Enterprise 6x10'6 Unmonitored concentration in extractant, transfer to unsafe geometry vessel.

12/13n78 Siberian Chemical 3x10's Unmonitored transfer of Pu metal into Combine storage container.

3-103 NUREG/CR-64 10

3.4.1.4 Large Storage Arrays Limited experimental data exists for large storage array systems of the type found ini fuel cycle facilities.

Use of the reactor excursion data assumes development of a flooded condition and a bounding estimate of total number of fissions of 5x10RI.

As in the case of moderated, reflected metal systems, the event is expected to be characterized by a single burst with little or no plateau period.

3.4.2 Nuclear Criticality Accident Source Term Estimates Estimation,f the potential impacts of a nuclear criticality accident involves estimation of the total amounts of radioactive material and energy generated in the event and consideration of the fraction of this material and energy that escapes from the facility. This section presents a method for estimation of criticality accident source terms using a uranium solution criticality as an example. The method is representative of the technical approach that may be applied to evaluation of generic criticality events. Methods for estimation of airborne release fractions and respirable fractions for release modes characterizing criticality accidents are also discussed.

Estimation of criticality accident source terms begins with development of a conceptualization of the physical system that retains the important characteristics of the system, but which allows application of mathematical models and analytical tools to predict the behavior of the system. The analysis aims to bound potential impacts rather than predict system behavior in great detail. To this end, the historically observed behavior of solution criticalities as involving an initial burst followed by an extended plateau is adopted as describing the event. The plateau period is represented as comprising a series of pulses each with total number of fissions less than the initial burst. Following reported accident events, the initial burst is taken to be lxlIO fissions, with 47 subsequent bursts each involving 1.92x10' 7 fissions, for a total of lx 10" fissions. Each burst lasts 5 seconds and is separated from the preceding burst by 10 minutes, yielding a total accident duration c, 8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br />. The example consists of a tank containing a solution of U0 2(NO3)2 enriched to 4 per cent in the U-235 isotope. The concentration of uranium is taken to be 400 g/9 (- 25 lb/ft3) and the total volume is 400 1 (- 0.011 ft3). The actual geometry of the system is approximated by a sphere and the physical state is represented as a homogeneous liquid mixture.

For the example, the analysis may be separated into estimation of fission product generation rates and estimation of gamma ray and neutron currents escaping the system. The fission product generation rates are estimated using the ORIGEN2 computer code (ORNL 1989). ORIGEN2 is a point-depletion and radioactive decay computer code that can be used to simulate nuclear reactor processes. The fission product yields for important isotopes predicted using the code are presented in Table 3-14.

Because ORIGEN2 does not provide output data on prompt gamma and neutron generation rates, a supporting model is needed. The gamma ray generation rate and energy distribution are adopted from power reactor operating data. The prompt photon production rate is reported as 7.03 photons per fission, with the energy spectra presented in Table 3-15. For the example case, a total of 7x10'9 prompt photons would be produced over the 8-hour period. The total production could be conservatively adopted as the release rate or the source could be represented as uniformly distributed through the spherical, aqueous system (with a radius of 45.7 cm (18 in)) and a shielding code used to estimate the rate and energy distribution of photons escaping the system.

A similar approach could be used to evaluate escape of prompt neutrons. For U-235, approximately 2.4 prompt neutrons are produced for each fission, yielding a total of 2.4x1i0' prompt neutrons for the entire NUREG/CR-6410 3-104

Table 3-14. Radioactivity Generated In a Uranium Solution Criticality Accident Nuclide Half-life Radioactivity, Bq (Ci)

Average Decay Energy (MeV) 0-0.5 hr 0.5-8 hr Total y

p Kr-83m 1.8 hr 7.4E11' 4.6E12 5.4E12 2.7E-3 2.9E-2 (20)

(130)

(150)

Kr-85m 4.5 hr 4.5E11 2.8E12 3.3E12 1.6E-1 26E-1 (12)

(77)

(89)

Kr-85 10.7 hr 6.7E4 4,1E5 4.8E5 2.2E-3 2.5E-1 (1.8E-6)

(L.IE-5)

(1.3E-5)

Kr-87 76.3 min 5.5E12 3.4E13 4.0E13 7.9E-1 1.3E0 (150)

(920)

(1070)

Kr-88 2.8 hr 3.4E12 2.1E13 2.4E13 1.9E0 3.6E-1 (91)

(570)

(660)

Kr-89 3.2 min 2.3E14 1.5El 5 1.7El5 1.6E0 1.3E0 (6300)

(40000)

(46000)

Nuclide Half-life Radioactivity, Bq (CI)

Average Decay Energy (MeV) 0-0.5 hr 0.5.8 hr Total 1 3 Sr-91 9.5 hr 1.6E12 1.0E13 1.2E13 6.9E-1 6.6E-1 (44)

(280)

(320)

Sr-92 2.7 hr 5.9E12 3.7E13 4.2E13 1.3EO 2.OE-1 (160)

(990)

(1200)

Ru-106 368 day 1.2E8 7.4E8 7.4E8 L.OE-2 (3.3E-3)

(2.OE-2)

(2.OE-2)

Cs-137 30.0 yr 6.3E7 3.7E8 3.7E8 1.9E-1 (1.7E-3)

(1.OE-2)

(1.OE-2)

Ba-139 82.7 min 1.3E13 7.8E13 9.1E13 4.3E-2 9.OE-1 (340)

(2100)

(2400)

Ba-140 12.7 day 5.5E10 3.5E11 4.1El I 1.8E-1 3.IE-1 (1.5)

(9.5) 011)

Ce-143 33.0 hr 5.2E11 3.2E12 3.7E12 2.8E-1 4.3E-1 (14)

(87)

(100) 3-105 NUREG/CR-64 10

O Table 3-14. Radioactivity Generated In a Uranium Solution Criticality Accident (Continued)

Nuclide Xe-133 Xe-133m Xe-135 Xe-135m Xe-137 Xe-138 Half-life 5.2 day 2.2 day 9.1 hr 15.3 rain 3.8 rain 14.2 irin Radioactivity, Bq (C0) 0-0.5 hr 0.5-8 hr Total 1.4E7 8.5E7 1.OE8 (3.7E-4)

(2.3E-3)

(2.7E-3) 9.6E7 6.3E8 7.0E8 (2.6E-3)

(1.7E-2)

(1.9E-2) 2.7E10 1.7E11 1.9E 1I (7.3E-1)

(4.5)

(5.2) 1.7E12 1.OE13 1.2E13 (45)

(280)

(330) 1.2E14 7.6E14 8.8E14 (3300)

(21000)

(24000) 5.2E13 3.2E14 3.7E14 (1400)

(8700)

(10000)

Average Decay Energy (MeV) 4.6E-2 1.4E-1 4.1E-2 1.9E-I 2.5E-1 4.3E-1 1.6E-1 1.1EQ 3.2E-l1 9.8E-2 1.8E0 6.7E-1 Nuclide 1-131 1-132 1-133 1-134 1-135 Half-life 8.0 day 2.3 hr 20.8 hr 52.6 min 6.6 hr Radioactivity, Bq (Ci) 0-0.5 hr 0.5-8 hr Total 3.7E10 2.3E11 2.7Ell (1.0)

(6.3)

(7.3) 5.2E12 3.2E13 3.8E13 (140)

(880)

(1,000) 8.6EII 5.4E12 6.2E12 (23)

(150)

(170) 2.1E13 1.3E14 1.5E14 (570%

(3,600)

(4,200) 2.6E12 1.6E13 1.9E13 (69)

(430)

(500)

Average Decay Energy (MeV) 3.8E-1 1.9E-1 2.2E0 4.9E-1 6.OE-I 4.1E-1 2.6E0 6.2E-1 1.6E0 3.6E-1

'exponential notatioun,7.4E1 I =7.4x IV' NUREG/CR-6410 3-106

Table 3-15. Prompt Fission Gamma-Ray Spectra*

E N(E)

(MEV) y's/fission 0.5 1.0 1.5 2.0 2.5 3.0 3.5 3.1 1.9 0.84 0.55 0.29 0.15 0.062 Table 3.16. Prompt Neutron Energy Spectrum Energy Fraction of (MeV)

Neutrons 0.0-0.2 0.038 0.2-0.4 0.061 0.4-0.6 0.069 0.6-0.8 0.071 0.8-1.0 0.071 1.0-1.2 0.068 1.2-1.4 0.064 1.4-1.6 0.059 1.6-1.8 0.055 1.8-2.0 0.050 2.0-2.2 0.046 2.2-2.4 0.041 2.4-2.6 0.037 2.6-2.8 0.032 2.8-3.0 0.030 3.0-3.2 0.026 3.2-3.4 0.023 3.4-3.6 0.020 3.6-3.8 0.018 4.0 0.065 4.5 0.024 5.0 0.019 5.5 0.017 6.0 0.007

ý65 0,004

• frm ANL 1963 3.8-4.0 4.0-5.0 5.0-6.0 6.0-8.0 8.0-10.0

>10.0 0.016 0.054 0.027 0.019 0.004 0.001 criticality event. The energy distribution reported (LaMarsh 1983) for these neutron is summarized in Table 3-16. As in the case of the prompt gamma rays, the prompt neutron source could conservatively be used as the released source or the source could be uniformly distributed through the spherical system, with the rate and energy distribution of neutrons escaping the system estimated using a shielding code.

In the case of neutrons, a third analytical approach could be applied. A single-group thermal diffusion model could be used to model the neutron flux, with the escape fraction estimated as the integral over the surface of the system of the product of neutron diffusivity and gradient of neutron flux. This approach yields a total escape of 2.7 xl07 neutrons during the 8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> of the criticality event. Application of these concepts is further illustrated in the sample problem discussed in Section 9 of Appendix D.

3-107 NUREG/CR-64 10

.9 3.4.3 Estimation of ARF and RF for Nuclear Criticality Accidents Estimation of the airborne source term resulting from an inadvertent nuclear criticality (INC) is based upon the same 5-factor formula used for other events. The evaluation of the various components may have expanded definitions. Some considerations in the evaluation of the five components are:

MAR. The MAR is the radionuclide inventory that can be at risk during the postulated event. Unlike the MAR for most events, where the MAR is a fixed value (typically the entire radionuclide inventory within some fixed physical boundary), the MAR for an INC depends on the fissile/fissionable materials involved and the total fission yield. As discussed previously, the inventory can be estimated by use of a computer code such as ORIGEN2, which evaluates the fission product (FP) inventory generated by a specific fissile material (i.e., "U, "SU, 1"Pu) during an nuclear excursion. In some instances (e.g., acid solutions of SNF and other materials that contain a mixture of fissile and fissionable materials), the other fissile and fissionable materials can fission (depending on duration of excursion and neutronics of the situation) and generate their own spectrum of FPs. Typically, the duration of the excursion is short, and contributions firom this source are ignored, but it should be borne in mind that even code values are approximations, with some level of uncertainty depending on the conditions.

DR. The DR reflects the fraction of the MAR that is actually impacted by the specific event postulated.

As an example, a fire may range from a small fire with a limited amount of fuel or oxygen availability affecting only a fraction of the MAR, to a large fire that affects all of the contents within the physical boundaries that prevent its propagation. For INCs, the DR is typically assumed to be 1.0.

ARM. Because of the various physical forms of the radionuclides (gases, vapor, non-volatile) and the matrix (solutions, metal, powders, irradiated compacted ceramic oxides [SNF] and cermets), there is more than a single ARF for the airborne release resulting from an INC. The ARFs for various physical forms of the radionuclides and matrix are discussed in greater detail below.

MF. The RF is defined as the fraction of the ARF that is in the particle size range DAED 10 micrometers and less. Since noncondensible gases and vapors (materials airborne in the gaseous state) are not particles, the RF applies gn& to materials that are made airborne as particles (non-volatile compounds in solid or liquid form).

LPF. There are many LPFs possible during the transport of materials made airborne during an event from the point of origination to release from the facility-atmosphere interface. The LPFs may be expressed as a single value that represents the summation of all LPFs, or that represents the single LPF phenomenon considered. In most event analyses, the LPF represents phenomena that occur after the radionuclides are made airborne in some confinement. In the case of INCs, there is one category of matrix (large storage arrays), with one class of materials, clad, (i.e., SNF, "pits"), where an LPF may exist at the point of origination.

For SNF, the clad is designed for operation at high temperatures and pressure found in nuclear reactors, and those that have not failed in this environment have a reasonable expectation of resisting the temperature and additional internal pressure generated by an INC. No airborne release is anticipated from intact, clad, SNF in an INC. The SNFs that have already failed are assumed to have released their "gap" inventory of FPs generated by irradiation in the reactor, and the same fraction of the INC-generated inventory would be released.

NUREG/CR-6410 3-108

4, 4

"Pits" (weapons-grade plutonium hollow metal spheres encased into a welded and tested stainless clad with and without a small-diameter tube leading into the internal cavity) also offer some degree of resistance to airborne release of the event-generated FPs. If the calculated metal temperature induced by the energy from the INC exceeds the melting point of the weld material (- 423 K 302 'F), the encased Pu would be exposed to air. At these temperatures, air oxidation of Pu is slow. If the temperatures exceed the ignition temperature (the temperature at which a self-sustained oxidation reaction is initiated, the metal temperature would increase to 1273 to 1373 K (1832 to 2012 *F), failing the stainless steel clad); all of the metal would be assumed to be oxidized; and the ARF and RF for Pu metal under thermal stress (ARF 5E-4 and RF 0.5) would also apply.

3.4.3.1 Rules-of-Thumb for Total Fissions during INC Because of the complexity of the phenomenon and the many factors that should be considered to obtain an estimate of the total fissions generated by a specific event, "rules-of-thumb" provide users with a quick, bounding assumption for total fissions and other phenomena that may affect the total release of radionuclides.

a.

Liquid Systems

1. Large systems involving greater than 380 f (100 gal). The reaction terminates by evaporation of 100 liters (26 gal) of liquid: 1xlO19 total fissions (WxlO" initial burst [0.5 seconds], followed by 47 bursts of 1.9x10' 7 fissions at 10 min intervals). The fission product generated is dependent on the initial fissile materials involved and should be estimated by a code such as ORIGEN2 or equivalent. The aerosol generated by the boil-off of water suspends 5x10" of the salt content of the liquid evaporated (100 liters (26 gal) required to terminate the reaction).

2.

Small systems involving less than 380 f (100 gal). The reaction terminates by eructation of a portion of the liquid (release from a free-fall spill of liquid and suspension of a fraction of the liquid): lxl0t" total fissions in a 0.5-s burst. Reaction is terminated by eructation of liquid, and a free-fall spill release value should be applied urider this assumption. The fission products generated are dependent on the initial fissile materials involved and should be estimated by a code, such as ORIGEN2 or equivalent.

For any liquid system, the assumptions for fission product release are:

All the noble gases generated are released

" Most of the volatile iodine isotopes are retained in the liquid and 25 percent is released.

b. Solid Metal Systems - lxlO" total fissions. Fission product generated depending on initial fissile materials involved and should be estimated by a code such as ORIGEN2 or equivalent. For release assumption, see text.

c. Powder Systems - Ix 10*' total fissions. Fission product generated depending on initial fissile materials involved and should be estimated by a code such as ORIGEN2 or equivalent. For release assumptions, see text.

d. Large Storage Arrays - 1xl0°' total fissions. Fission product generated depending on initial fissile materials involved and should be estimated by a code such as ORIGEN2 or equivalent.

3-109 NUREG/CR-64 10

Tehnical Bais W

The value for the solid metal system and powder systems is based upon avalue that bounds most of the reported data for inadvertent nuclear criticalities that have occurred and are listed in Tables 3 "Summary of Known Accidental Criticality Excursions (a. Solution Systems, b. Metal Systems, and c.

Moderated Foil and Powder Systems (1945 to 1974)"; 3 "Accidents in Processing Plants"; 3 "Destructive Power Excursion Summary"; 3 "Inhomogeneous Water-Moderated Systems"; 3 "Miscellaneous Systems"; and 3 "Nuclear Criticality Accidents at Russian Industrial Facilities." All but one of the inadvertent nuclear criticalities that have occurred in processing facilities have involved solutions. Solids are more readily controlled because of their fixed physical configuration. Likewise, powders tend to flow and much more than the minimum critical mass stated in the standards (ANSI 1983) and other relevant documents (Clayton 1979; Thomas 1978) is required for criticality in the conditions and geometries encountered in processing plants. McLaughlin (1991) presented a table of "Criticality Accident Fission Yields" that was proposed for estimation of safe exclusion areas within work places (Table 3-17 "Criticality Accident Fission Yields"). The values are larger in many cases than those cited here.

The total fission value for large storage arrays is based upon the data presented of the events that have occurred, consideration of the yield from reactor excursions, and Woodcock's estimates shown in McLaughlin's table. No estimates are made for the fission product release.

3.4.3.2 Scenario Assumptions and ARFs and RFs Solution Systems

1. Scenario Assumptions. As discussed in Subsection 3.4.2.1, "Nuclear Criticality Accident Source Term Estimation Methods," the assumed total fission yield from the event in a spherical system 45.7 cm (18 in) in radius is an initial burst of lxl0'8 fissions, followed by 47 5-sec bursts of 1.92x10"7 fissions each 10 minutes, for a total of lx10"9 total fissions. The volume of a sphere with a radius of 45.7 cm is approximately 400 liters (- 100 gal). The "rule-of-thumb" for the total fission yield is based on the estimated total fission from the INCs that have occurred in solutions and provides a reasonable upper bound value for events other than those that occur in very large volumes that greatly exceed the critical mass.

2. ARFs and RFs. The estimated inventories from an INC involving `'U in solution is shown in Table 3-14, "Radioactivity Generated in a Uranium Solution Criticality Accident." The physical forms of radionuclides generated (shown in Table 3-14) are those typically involved (although the inventories differ) and include noncondensible gas (noble gases), volatiles (iodine in heated, acid solution), and non-volatiles (all remaining radionuclides that are dissolved in the acid solution). Thus, ARFs are required for the noble gases and iodine and ARFs and RFs are required for the non-volatile materials generated and present in the solution (the fissile material and other non-volatiles present in dissolved SNF).

Noble gases - All noble gases present are assumed to be released during the generation and subsequent boiling of the liquid. The ARF assigned is 1E+0 (unity).

NUREG/CR-6410 3-110

Table 3-17. Criticality Accident Fission Yields (McLaushlin 1991)

System WnitiaIBurtYield Total Yield Solutions under 100 gal lxlo'7 3x10 1 R

(0.38 m3)

Solutions over 100 gal IxlO" 3x10' 9 (0.38 M3)

Liquid/powder 3xl0 3x1020 Liquid/metal pieces 3x100 11XO19 Solid uranium 3x 10" 3x 10'9 Solid plutonium I xl011 lxiO Large storage arrays below None 1x 1019 prompt critical Large storage arrays above 3 x10 22 3x 1022 prompt critical Iodine - All iodine isotopes are assumed to be quantitatively released during the boiling of the acid solution. This includes all iodine isotopes that may already be in solution (e.g., 1291 and "'tI in dissolved SNF). The ARF assigned is lE+0 (unity). Iodine is a highly reactive chemical material that will react with materials (structural, airborne inert particles, engineered exhaust treatment components such as filters and absorbers) encountered along pathway to release. The typical assumption is that an upper bound of 0.25 of the iodine released from the liquid escapes the facility and is released into the atmosphere.

Non-Volatile - The suspension mechanism for non-volatiles in solution is the formation and suspension of liquid droplets from film-break up. Films are formed by the generation of bubbles of gas or vapor as they rise to the surface thinning the upper film of the bubble until the bubbles burst.

The experimentally measured upper bound value for the ARF is 2E-3, with no measured RF. Since the boiling is assumed to reduce the volume of solution to a level that is no longer a critical configuration, not all the liquid is evaporated in the process. For small volumes, there is a reasonable expectation that the initial burst of lxlO0l fissions should result in a sufficient loss of volume to terminate the criticality excursion. Thus, it is assumed that less than 25 percent of the volume is evaporated and that the applicable ARF is [2E-3][0.25] or 5E-4 with an RF of 1.0.

The ARF and RF values are applicable to All non-volatile compounds in the liquid, including the fissile material and non-volatiles present in the reacting solution and all the INC-generated FPs. As shown in Table 3-14, for an INC involving 11U, the highest inventory of non-volatile FP generated by the excursion is 1200 Ci of 92Sr.

3-111 NUREG/CR-6410

-)

i Scenario Assumptions. An INC involving ametal would result in FPs formed throughout the metal matrix. Unless energy levels are sufficient to soften the metal, only the Fts on the metal surface are likely to be released. Restrepo (1992) reviewed the literature on FP release from failed fuel elements heated to a temperature that resulted in fuel slumping. He divided the elements into categories that appear to behave similarly under these conditions. The ARFs derived are shown in Table 3-18, "Release Fractions for Various Chemical Classes from Heated Spent Fuel". The values represent the release estimated for unclad SNF heated to temperatures exceeding those anticipated for INCs and are easily upper-bound values for these materials from INCs.

Powder Systems I. Scenario Assumptions. As with a metal matrix, an INC involving fissile materials in powder form would result in the formation of FPs within the matrix of the material involved. The tendency is to favor formation near the surface region where the neutron flux is greater. With the larger surface area per mass of powder than for a metal, the values applicable to metal are not relevant.

2. ARFs and RFs. The arface to mass ratio for a powder is dependent on both its particle size distribution and shape. Both are unknown and the upper-bound values are conservatively assumed to be:

Noble gases - ARF IE+O, RF NA Iodine - ARF I E+O, RF NA Non-volatile - [based on the suspension during the heating of an chemically, non-reactive compound from,(DOE 1994, p. 4-57)] ARF 6E-3, RF 0,01.

Large Storage Arrays

1. Scenario Assumptions. This category of material represents a very large inventory of fissile materials storage in a geometrically favorable configuration. It is impacted by an event that defeats the system to produce a geometrically unfavorable configuration.

There are two relevant classes of materials:

1. Metal Clad Material - SNF, "pits"

2. Unclad - metal, powders, etc.

The anticipated behavior of the two classes of Large Storage Array materials was discussed in Subsection 3.4.3, LPF. The amount released to the ambient atmosphere is the product of the ARF & RF, if applicable, and the LPF.

2. ARFs and RFs for INCs Involving Large Storage Arrays. For the unclad materials, the ARFs and RFs for the material forms are applicable.

For clad materials, the integrity of the clad during and following the excursion determines the release. The amount of material released through the clad may be considered LPF1 (Leak Path Factor, initial - the fraction of material released within the clad that penetrates the clad).

NUREG/CR-6410 3-112

There is more than one form of SNF. The common forms are a compacted, sintered, ceramic oxide pellet clad in zircaloy or stainless steel (commercial nuclear fuel); uranium metal clad in aluminum or zircaloy; enriched uranium metal-aluminum alloy; cermets, etc. There has been experimental measurements of FP release during fuel failure from heating, and metal alloy and cermet targets. The derivation of ARFs from ceramic oxide fuel heated to slumping has been covered in Subsection 3.4.3.1.b ("Metals"), and these are applicable for this class of clad material.

Table 3-18. Release Fraction for Various Chemical Classes from Heated Spent Fuel

(_Restrei

_o1991)

Group Group Nae.

Rep. Ele.

Elemens In Group ARF 1

Noble Gases Xe Xe, Kr, He, Ne, Ar, Rn, H 5E-1 2

Alkali Metals Cs Cs, Rb, Li, K, Fr, Na 2E-l 3

Alkali Earths Ba Ba, Sr, Mg, Ca, Ra, Be 3E-2 4

Halogens I

I, F, Cl, Br, At 5E-2 5

Chalogens Te Te, S, Se, 0, Po, N 7E-2 6

Platinoids Ru

.Ru, Rh, Pd, Os, Ir, Pt, Au, Ni 2E-3 7

Transition Mo Mo, V, Cr, Fe, Co, Mn, Nb, Tc 3E-2 Metals 8

Tetravalent Ce Ce, Ti, Zr, Hf, Th, Pa, U, Np, 4E-4 Pu 9

Trivalent La La, Al, Sc, Y, Ac, Pr, Nd, Pm, 6E-4 Sm, Eu, Gd, Th, Dy, Ho, Er,

Tmn, Yb, Lu, Am, Bk, Cf 10 Main Group I Cd Cd, Hg, Zn, As, Sb, Pd, TI, Bi 4E-3 11 Main Group II Sn Sn, Ca, In, Ag 4E-3 12 Boron B

B, Si, P, C 6E-4 The "'perimental data for release from metals and cermets were analyzed in DOE (December 1994). The ARFs are divided into "instantaneous" (recovered during a 2-mnin collection period) and "total" (recovered during a 60-min collection period). Only values for three elemental forms were given:

"Instantaneous" "Total" Cesium Iodine Tellurium 0.06 0.8 0.00 0.09 0.9 0.007 3-113 NUREG/CR-6410

• . I-I V Although the other non-volatile radionuclides are not specifically listed, the values for tellurium, which can be volatile under some conditions, can be applied as upper bounds.

3.5 Chernicals The purpose of this section is to describe how to calculate the characteristics of accidental releases of hazardous chemicals such as uranium hexafluoride, hydrogen fluoride, ammonia, chlorine, and other chemicals commonly found at nuclear fuel cycle facilities. The characteristics of the chemical source term include rate of release, temperature, momentum, orientation, height, and aerosol content. These quantities are required for atmospheric dispersion modeling, which is discussed in Chapter 5 of this Handbook.

3.5.1 Identification of a Representative Range of Source Terms No single atmospheric dispersion model or source term will suffice for the range of scenarios possible at the fuel cycle facilities. At the same time, it may not be necessary to develop or use highly sophisticated models that take into account every possible detail about the source term. Certain aspects of the source term need not be fully quantified, if they can be shown to have little or no impact on the final risk estimate.

Hence, the selection of the appropriate source term is as much an art as a science. This section provides guidance on the selection of the appropriate source term for the types of accident scenarios involving chemicals possible at the fuel cycle facilities by providing a broad overview of the range of source terms.

For modeling details of each type of source term described herein, the reader is directed to Appendix B of this Handbook.

3.5.1.1 Liquid below Its Boiling Point If spilled accidentally, liquids with boiling points that are well above ambient temperature will form a pool on the ground and evaporate slowly at a rate determined by the ambient temperature, the windspeed and the area of the spill. No aerosolization is expected in such a spillage. Hydrofluoric acid (normal boiling point 293 K (68 *F) is an example of a liquid that can be spilled below its boiling point if the spill occurs from a storage tank at an ambient temperature that is less than 293 K (68 OF).

If the spill initially occurs from a severed line, then Bernoulli's equation can be used to estimate the flow rate onto the ground (see equation B. I in Appendix B). Bernoulli's equation takes into account the influence of static liquid pressure head, and the influence of vapor pressure above the liquid level.

After a spill, the methods developed for estimating the evaporation rate of slowly evaporating liquids from pools can be used. These methods assume that heat transfer from the underlying surface of the pool is rapid, and is not the limiting factor for evaporation. The rate of evaporation is limited by the rate of mass transfer across a stagnant film of air at the surface of the pool. The method for predicting the evaporation rate is described in Section B.2. 1, equations B.2, B.3, and B.4 of Appendix B. An example of a calculation of an evaporating pool of HF at a temperature of 280 K (44 *F) is given in Sample Problem No. 5.

3.5.1.2 Refrigerated Liquid in a Vessel The method used to estimate the release rate of a liquid below its boiling point (Bemoulli's equation), can also be used to estimate the release rate of a refrigerated liquid. Examples of refrigerated liquids include freon and ammonia. In the present context, refrigerated liquids have boiling points that are well below NUREG/CR-64 10 3-114