ML20216B700

Forwards Compliance Evaluation Rept Prepared to Support Amend of Coc GDP-2.Notice of Amend,Forwarded to Ofc of Fr for Publication,Also Encl

ML20216B700
Person / Time
Site:	Portsmouth Gaseous Diffusion Plant
Issue date:	09/02/1997
From:	Yawar Faraz NRC OFFICE OF NUCLEAR MATERIAL SAFETY & SAFEGUARDS (NMSS)
To:	John Miller UNITED STATES ENRICHMENT CORP. (USEC)
Shared Package
ML20216B706	List: ML20216B700 ML20216B716
References
TAC-L32032, NUDOCS 9709080130
Download: ML20216B700 (15)

Text

_ _ _ _ . _ ___ _ .._.- _ __. __ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ . . _ . _ _ . _

t s ,

... September 2, 1997 l

.. o l ,

Mr. James H. Miller i

Vice President, Production O. S. Enrichment Corporation 2 Democracy Center  !

6903 Rockledge Drive Bethesda, MD 20817 l

SUBJECT:

CERTIFICATE AMENDMENT REQUEST PORTSMOUTH GASEOUS DIFFUSION i PLANT ERP SCALE PIT RASCHIG RINGS MINIMUM DEPTH REDUCTION (TAC l NO. L32032) ,

i

Dear Mr. Miller:

Enclosed is a copy of the Compliance Evaluation Report prepared to support the i amendment of Certificate of Compliance GDP 2. A copy of the Notice of Amendment, t

which has been forwarded to the Office of the Federal Register for publication, is also i enclosed. This notice provides the opportunity for the public to petition for review of the i decision in accordance with 10 CFR Part 76, Subpart C. Final action on your amendment  !

i request will not be taken until af ter the time allowed for requesting review of the Director's f

l Decision is over, if you have any questions regarding this action, I can be reached at (301) I 415 8113.

l l

Sincerely, 1O' Original Signed By I

(

Yawar H. Faraz, Project Manager .

"r 9709080130 970902 Enrichment Section PDR ADOCK 07007002 Special Projects Branch C PDR Division of Fuel Cycle Safety and Safeguards, NMSS Docket 70 7002 Certificate GDP 2 h,, h((kh(h ] hhfy cc: Mr. Steven A. Toelle, USEC

Enclosures:

1. Compliance Evaluation Report l, l ll*'

2. Notice of Amendment l' lll'lil'll DISTRiauIl0N: (Control No.1305) '

Dechet 70 7002:. NRC Fee Conto, PUDLIC - KO'Bnen. RHI CCom,RHI NMS$ t/f FCSS t/f I

$PB t/f oGC . PHiland Mill DHeritand RHI g gk7 JDavia. FCOB OFC GPS b bP8 SPB t $4 NAw - vr.,N,D2 No.rii., nMire Nion DATE '80f /97 8tM 197 81 97 8 197 / 19 7 # /97 C = COVER E = COVER & ENCLOSURE N = 900 COPY OFFICIAL RtCOMO COPY t

V, v t. b .= 1

'e .

@ #80 g yo k UNITED STATES

• } NUCLEAR REGULATORY COMMISSION '

• WAsMNQToN. J C. Poth401

$,4 . , , , , */ Septernber 2,1997 DOCKET: 70 7002 CERTIFICATE HOLDER: United States Enrichment Corporation ,

Portsmouth Gaseous Dif fusion Plant I g.

Portsmouth, Ohio ,

SUBJECT:

COMPLIANCE EVALUATION REPORT: APPLICATION DATED MAY 16,1997, ERP SCALE PIT RASCHIG RINGS DEPTH BACKGROUND On May 16,1997, United States Enrichment Corporation (USEC) submitted a request to reduce the minimum depth of Raschig rings from 12 inches to 6 inches in Scale Pits 1 A and 2 in the Extended Range Product (ERP) f acility at the Portsmouth Gaseous Diffusion Plant (PORTS). The minimum depth is specified in Technical Safety Requirement (TSR) 2.5.4.4 entitled " Scale Pit Raschig Rings."

The ERP f acility is used to withdraw enriched UF from the cascade and fill cylinders which are placed on scales at the withdrawal stations. Neutron poison in the form of borosilicate glass Raschig rings are used in scale pits of the product withdrawal stations at PORTS to increase the mass of uranium and moderator necessary for a criticality to occur. The proposed certificate amendment request is required to allow proper operation of the scale mechanism at the ERP 1 A station. USEC has also requested reduction of the required depth of Raschig rings at the other stations to maintain consistency of administrative control on this neutron poison parameter. Scales perform the safety function of measuring cylinder weight to prevent overfilling. When heated, an overfilled cylinder could rupture and release UF.,

The Design Feature (DF) requirement of TSR 2.5.4.4, as described in the USEC certificate application, states:

"DF: . Scale pits shall contain Borosilicate glass Raschig rings to a nominal depth of 12 inches at ERP (positions 1 A and 2),6 inches at ERP (position 18) and 6 inches at LAW / TAILS."

USEC has proposed to revise the DF requirement to state:

1

. _ _ _ _ _ . _ _ _ _ _ __ _ _ _ _ _ . . _ m _ _ . _ _ _ _

f g ,

l "DF: ERP, LAW and Tails scale pits shall contain Borosilicate glass Raschig rings '

to a m;nimum depth of 6 inches "

The staff independently performed a safety review of the proposed action. The NRC staff review resulted in an information request (RAll from Yawar H. Farar to James H. Miller dated June 23,1997 (RAll). RAll requested USEC to determine whether actuation of the sprinkler system could be caused by the heat of reaction of accidentally released UF with water vapor and other substances in ERP. USEC's response to RAli was from James H. ,

Miller to Carl J. Paperiello dated July 7,1997 (RSP1). It concluded that "the potential release of UF. In the ERP f acility resulting in a sprinkler actuation due to the exothermic reaction of UF. and atmospheric moisture within the ERP facility is not credible." A review  ;

of RSP1 tesulted in additional questions regarding the assumptions and data used in the l calculations. Responsos to these questions were provided by Ed Wagner of LMUS in a 1

phone conversation with Yawar Faraz on July 15,1997. These responses are documented in a telecon note dated July 10,1997. ~

DISr.USSION ERP is one of three permanently established f acilities wnich withdraw uranium in the form of UF from the cascade, it has three withdrawal stations, namely Stations 1 A,18 and 2, l and is located in the northeast corner of the X 326 Process Building. Enriched UF. product l ls withdrawn from the cascade into cylinders in liquid form at these three stations. To

accomplish this, centrifugal withdrawal compressors (two in series within each compression loop) compress the gas stream to a pressure (30 to 35 psia) and temperature '

(high enough to maintain UF,in vapor phase) above the triple point and then condensers cool and condense the UF. vapor, Accumulators are provided to accumulate liquid UF.

l while cylinder filling is interrupted. See Figure 1. The compressors are located on the second floor of the process building. The accumulators, associated piping, UF condensers, and valves are located on the mezzanine level between floors. Withdrawal stations t (manifolds, cylinder scales and cylinder carts) are located on the ground floor. Each withdrawal station has a floor mounted scale on which a cart mounted cylinder is placed, l Operating mechanisms of each scale are housed in a pit directly below the product cylinder.

The scale pits at ERP 1 A and ERP 2 are approximately 12 feet by 8.5 feet, Scale pit ERP-1 A has a depth of 72 73 inches and scale pit ERP 2 has a depth of 50 51 inches, At a minimum, each scale pit is filled to a six inch depth with a neutron poison (Raschig rings) and covered by steel floor plates, All equipment and piping leading from the cascade to the

( cylinder are " safe geometry" for the allowed enrichment assays.

Each cylinder to be filled is connected to the fill manifold via a pigtail assembly. The pigtail attachment area is monitored by two Pyrotronics smoke detector heads which when actuated by a UF release, willisolate the cylinder pigtail by closi.)g the air-operated safety valves on both the cylindar and on the fill manifold, Euh withdrawalloop also has a High Pressure Venting (HPV) circuitry which is manually a:tuated on indications of UF. leakage to relieve the system pressure, if the HPV circuitry is actuated, a control valve in the compression loop automatically opens and vents process gas back to the cascade via the vent return header, the flow control valve (s) to the condenser close, the condenser vent valve closes, the air operated manifold isolation valve on the cylinder pigtail closes, and the 2

i L - --- - . -

's s .

first-stage compressor suction valve closes. The HPV is also auto wtically activated on a high compressor discharge pressure indication.

The staff used accepted industry manuals as Oiven in References 2. 3 and 4 to evaluate the stated USEC hmits. For cases where tabulated industry data was limited or where the postulated scenanos represented a more complex system than could be handled by reference material, the computer code KENO.V.a was used.' This code is a multi group, three-dimensional Monte Carlo program that is widely used for criticality safety assessments in the nuclear industry. The cross section hbrary used with this code was the 27-group ENDF/B IV neutron cross section library, maintained as a database within the SCALE system.' The 27-group library was collapsed from broad group data provided in the Evaluated Nuclear Data File /8 Version IV.' This broad group hbrary was developed pnmanly for criticality safety analyn and has undergone extensive evaluatirn.

The majority of the staf f's calculations used UO,F, (uranyl fluoride - the hydrolysis product of UFe) to determine the multiplication of the system since a review of cntical experimental data in Reference 2 shows that the minimum homo 0eneous critical mass for UO,F,is smaller than UFe . This is due to the fact that oxygen present m uranyl fluoride is a more ef fective moderator than fluonne. Although a release of UFe would result in a r*.ture of UF, vapor, ufo solid, and the hydrolysis products UO,F, solid and HF vapor hmited by the entrained water in air, the analysis conservatively assumed that a leak would immediately flash to hydrated UO,F,.

SIAEFEALUATION ANDESULIS Twenty one major releases of UF have occurred at the three gaseous diffusion plants (PGDP, PORTS, and K 25) during the 18 year period between 1961 and 1978." Twelve of these releases occurred at PORTS and more than half of all major releases have involved pigtail connections. Releases involving liquid UF, are considered the most severe because of the rapid release rate and because they are difficult to contain. All release scenarios at the withdrawal stations involve hquid U', release, and therefoia, control mechanisms must be carefully evaluated to ensure that they perform their intended function as desi0ned.

Chemically, UF is a highly reactive substance. At normal ambient temperatures, UO,F2 sohd and HF vapor results from UF e hydrolysis. This hydrolysis reaction is a fast, hi hly 0

exothermic reaction going to completion at a rate hmited by the rate at which water vapor is entrained in air. The remaining unreacted UF is present as a concentration of vapor and sohd particulate mass. The reaction with moist air is given by:

UF, + 2H,0 - UO 2F, + 4HF + Heat (1)

UO 2 F2 + moist air -- UO,F(nH 2O (2) 3

i . . .

4 The hydrolysis product, uranyl fluoride, is hygroscopic and will continue to absorb moisture from the air, as given by equation 2, untilit reaches its saturation limit (approximately H/U=16).' At moderation ratios much below H/U=16, the system exists as a " slow flowing" material much like the consistency of syrup; at moderation ratios below H/U=9, the system becomes more like -

toothpaste; and below H/U=4,it is a mixture of solid hydrated uranyl fluoride compounds.

ADEQUACY OF CONTINGENCIES USEC commits to controlling mass and moderation for this operation. The Raschig rings are used as a tertiary control and contribute to the effectiveness of the mass and mooeration controls by increasing the quantities of uranium mass and moderator that would have to reach the pits for a enticahty to occur.

Mass Control Mass controlis based on prevention of UF releases as described in the accident analysis section of the SAR.

Piotail Failure USEC concluded that the only credible scenario that cou'd result in a large release of UF, above the scale pits is a pigtail rupture. According to the SAR, the Pigtail Line isolation System would limit the liquid UF. release to no more than 127.5 pounds. USEC conservatively assumed that all of the UF. is converted to UO,F, and that all of this material depohits in a scale pit. The resultant depth of the material, as determined by USEC, is 0.03 inches which is less than the 3.6 inch safe slab depth for 10% enriched material. The Raschig rings wero not included in this determination. However, the value of 0.03 inches is based on anhydrous UO,F, (" dry"); it is reasonable to assume that some hydration of UO,F, will occur over time which would increase the depth of material deposited in the scale pit. The staff determined that UO,F, fully hydrated to the saturation limit would occupy a slab depth of 0.12 inches which is still much less than the 3.2 inch critical slab depth for 10% enriched UO,F, at an H/U=16.8 Accumulator Failure - The staff also considers the failure of the accumulator and associated piping, which feeds UF. cylinders via gravity, to be a credible accident scenario Such a failure -

would release considerably more UF. than a pigtail rupture. Therefore, the staff determined the safety of an accumulator failure in which all of the released material transferred into a scale pit

-without Raschig rings present. The results of these calculations are given in Table 1.

4

i . .

Table 1: 2900 lb UF, Spill from ar. ERP Accumulator Loop into an Uncovered Scale Pit' cue 10 DESCRIPilON H/U Ken LiGMA E RP1 A 2918 Futenses Range Product Wthdreest 1 A Swe Pit 10 0 734 0 0039 I RP1 A 2912 (stensed Range Prod et Wfthdraw Y 1 A $cate Pit 12 0 620 0 0033

[RPtA 206 E siended Range Prosset Wthd<swal 1 A $cale Pd 0 0 4967 0 0030 LRP1 A 2004 Estended Range Product Wthdranal 1 A $cale P4 4 0 364 0 0029 ERP1B 2916 Estenced Range Product Wthdrawat iB $cate Pit 16 0 928 0 0037 ERPtB 2912 [rtended Range Produci Wths' anal 1B $c*e Pat 12 0 806 0 0038 CRP1B 2900 Estended Range Product Wthdranal ip $cate Pit 0 0 652 0 0034

[DP1B 2904 Estended Range Product Withdramat 1B Scre Pit 4 0 463 0 0024

• 1p. Enrched Vof ,. no boron pomon All cases in Table 1 demonstrate that a release from the withdrawal side can be tolerated. The analyses hypothetically assumed that enough moisture was available in the air to react with all of the UF, and continue to hydrate the resultant UO,F, product to the saturation lirait. No credit was taken for a manual activation of the HPV circuitry which would limit the release.

j Cylinder Failure - A cylinder release failure at any of the withdrawal stations could occur by three mechanisms: 1) a cylinder va!ve fa!!ure; 2) a passive structural failure: or 3) a pigtail failure in which the pigtailisolation system is unable to close the cylinder safety valve. Because operating methods and quality controllimits exist governing the inspection and certification of cylinders and cylinder valves, the probability of a cylinder or valve failing without outside interference is low. The pigtail line isolation system is designated a 'Q* safety system and receives quarterly surveillance tests to verify that the system willisolate the two safety valves.

USEC concludes that even a liquid cylinder passive failure (a very unlikely event) would not add enough uranium to the pit to cause a criticality, USEC's conclusion is based on postulated UF.

releases resulting in 15,000 pounds in ERP scale pits 1 A and 2 (dimensions 12 feet x 8.5 feet x

~4.5 feet) and 11,000 pounds in ERP 18 scale pit (dimensions 11 feet x 6.25 feet x -4 feet) which also bounds the LAW and Tails scale stations.

The staff performed computer calculations to determine the consequences of a full passive cylinder failure. The results of the calculations are given in Table 2 below. The ERP 1B scale pit was conservatively used in the calculations because its dimensions are more favorable to criticality. For all credible UF, release scenarios, the staff assumed the entire released amount to be transported into the pit. This is a very conservative assumption since almost half of the released UF, would immediately flash to vapor with the remainder becoming sol:.1. Also, during normal operation, the steel plates that cover the scale pits would restrict a large portion of the solid UF fraction from entering the scale pit. In addition, depending on the ambient temperature, a significant portion of the solid UF, could sublime to its gaseous phase.

5 i

- l

'e . .

,e Table 2: UF. Release and Settling in the ERP1B Scale Pit

• c... io riwwaiens wu Ra m,gReg risw. tas. 4ni Ur, neve s.a%

Depth (in) Hmght (cF.) Discharge (It>5 )

i RP1D 1610 00/, is 0 10 2776 0 779 0 0034 (RP1B itil UO/, 16 0 I t, 4105 0 97J 0 0031 iRP1B 1616 UO/, 16 0 16 4442 1 000 0 0032

( RPt B 'tBM6 UO/, 16 0 rutt 26000 1 369 0 0026

[RP1B UF6 Ur, - 0 40 6 26000 0 604 0 0023 LRP1B R26 UO/, 16 6 26 6926 0 613 0 0032

[RPiB R30 UO/, 16 6 30 7314 0648 0 0035

[RPiB R3$ 00/, , 16 6 35 8702 0 999 0 0037 LRP1B RR21 UO/, 4 6 21 2 11026 0 414 0 0019 (RPIB RR36 UOf, 4 6 36 19099 0 924 0 0032 ERP1BRR40 UO/, 4 6 40 2 26000 1 115 0 0032

' open pit (no top plate: 5% ennches mateenat As can be seen in the table, the multiplication of the system is significantly dependent on the H/U atom ratio of the uranyl fluoride hydrolysis product which directly correlates to the system slab height. As the H/U ratio increases, the effective quantity of reactant (UF6) that can be j tolerated is greatly reduced.

l The results demonstrate that a full passive 10 ton cylindet failure can be tolerated at en H!U ratio of 4 with a six inch Raschig ring depth (see case ERPiB.RR35). A full passive 14 ton cylinder failure can be tolerated only if it doesn't become moderated. However, it should be obvious that the H/U ratio is dependent upon the entrained water in the surrounding air.

According to USEC, after a leak is detected, the ventilation system to the withdrawal station is disengaged such that the volume of air in the room essentially becomes static, The staff a&cumed 30 pounds of water is available in this volume of air to react with the UF.. This would produce a relatively small quantity of UO,F2 (255.6 lbs.). The remainder of the release would remain as a concentration of UF. gas cnd particulate mass. Case ERP18.UF6 shows that the full contents of a 14 ton cylinder in the ERP1B pit as UF.would be very suberitical The small amount oi uranyl fluoride that would coexist with this UF, would not increase the predicted Keff by any appreciable amount.

Balchia Rina Failure - USEC states that the Raschig rings which are present in the pits increase the quantity of uranium mass that would have to reach the pits for a criticahty to occur.

However, the staff notes that in their amendment application, USEC did not consider the condition under which maintenance activities are conducted on a scale pit and the pit cover plate and neutron poison are removed for extended periods, According to USEC, operation of adjacent withdrawal stations while a scale is down for maintenance is an acceptable operating practice. Therefore, UP. releases must be considered for cases in which bare (i.e., no poison) scale pits have their cover plates removed and are open to moderator or fissile material 6

I

• e . .

]

, intrusion. However, the staff has already determined the safety of operations for which no boron poison is present as given in Tables 1 and 2 The calculations in Table 2 which included the Raschig rings considered the neutron poison as a homogeneous mixture evenly distnbuted with the uranyl fluoride that occupies the first six

, inches of the scale pit. Historical calculations completed by USEC also considered the borated glass as a homogeneous mixture. However, in the actual case,',he heterogeneity of the rings will reduce the effectiveness of this poison due to self shielding. Therefore, the staff attempted to quantify this reduced boron effectiveness. The results of this evaluation are given in Table 3 below. The calculated results were determined from an infinite slab of uranyl fluoride at an H!U atom ratio of 16. The material was either homogeneously distributed with the specified six inch depth of Raschig rings or was dispersed within a uniform lattice of Raschig rings. A layer of full density UO,F, hydrate was placed on top to simulate the geometric

] configuration of an actual release in the scale pits.

1 1

Table 3: Boron Effectiveness Determination j Case 40 ' 'OLUME MING Keff Keff Hedx tion Average FRACTION THICKNI$$ UN) homogeneous betrogeneous t%I Sigma 1 Baron 12!12H 0240 1.8 0.377 0439 14 0 002 1

! Boron N24 0 240 No Baron 1 059 + 0 003 I

As can be seen, the assumption of homogeneous distribution non conservatively reduces the l

! " actual" Keff by 14%. If we return to case ERP18.RR35 in Table 2 and adjust the effective multiplication factor by 14% (i.e.,0.924/[1-0.14]), we see that the predicted value would be critical. This indicates that the contents of a fuii 10-ton cylinder in ths r.cale pit cannot be tolerated under hydrated conditions. But again, this scenario is highly conservative because it assumes that the entire content of the cylinder ends up in the scale pit including the fractior' of the contents that flashes into vapor at the time of the release and because there would no' be enough water in the air to react with all of the UF., More interesting, if the rings are in place with a reduced boron content, the multiplication of the system dramatically rises (compare cases Boron.12 [ hetero) and Boron.N24). Although case N24 was run with a void fraction instead of modeling boron-deficient glass rings, the absorption characteristics of the other constituents are minor compared to that of boron 10 and the case serves to demonstrate the importance of maintaining the boron cont ,it of the rings.

Moderation Contro]

There are two sources of moderation that are of different levels of concern: 1) moisture contained in air; and 2) visible moisture that may exist and go undetected la the pits. A characteristic of low enriched uranium is that enrichments less than about 5.6 wt% "5U cannot be made enticalin the absence of a moderator. Thus, moderation controlis of specialinterest in these operations where characteristically 5 wt% materialis withdrawn.

In their response in RSP1 to NRC's RAli, USEC stated that based on a conservative calculation,17 kilograms of water vapor would be in ERP. USEC states in its evaluation that a 7

l I

4.--,$,4a+es-Gi. E 4 a5.44h6 -a,e D. - - - i-am- +-.e&4-a..A-Au,-m #2-4 i ..

T l minimum of 13 Kg (28.6 lbs) of water is required for a criticality to be possible for 10% ennched ,

UO,F, under optimum conditions. Later, USEC states that a minimum of 10 Kg or 22 poends of water is the minimum required for a criticahty to be possible for the same material. The staff investigated the apparent discrepancy and determined that 13.25 Kg (29 lbs) of water is needed for 10% enriched material." Five weight percent material requires a minimum of 22 6 Kg (49.72 lbs) of water.

~

However, completing a buckling conversion of a sphere of this material to a parallelepiped slab results in a critical depth of 19 cm. As previously discussed, the staff calculated that approximately 30 pounds of water under ideal conditions could or.:y produce 255.6 lbs of UO,F,. Even if this amount of UO,F, were considered optimally moderated as above, it would  ;

only occupy a slab depth of less than one centimeter (0.74 cm) which is clearly much less than the minimum slab depth of 19 cm. Furthermore, under normal conditions no fissile materialis expected to be uncontained within the withdrawal areas.

Of greater concern, is en inadverient introduction of moderator in the scale pits. Moderation in the pits could be from a fire suppression sprinkler discharge, leaks from RCW equipment located above the withdrawal stations, or other such similar sources. Moderation controlis based on the unlikeliness for any significant quantities of water to be found in the scale pits.

PORTS has instituted an administrative control to ir,spect the scale pits weekly for moisture.

Any water discovered is to be removed expediently. The staff agrees that it is unkkely that significant quantities of water would be found in the scale pits under normal operating conditions.

One final note, the moisture monitoring channels are 3 inch diameter tubing that run the full depth of each scale pit. The tubes are surrounded by concrete and do not contain any neutron poison. Although USEC does not discuss the nuclear safety of these cylinders if accidentally filled with fissile material, the staff determined that infinite cylinder diameters of 6 inches for the H/U of interest would be required for criticality of 93% enriched UO,F,8 Therefore, the staff determined that the 3 inch channels will be subentical for 10% enriched materials under accident conditions.

ADEQUACY OF CONTROL INDEPENDENCE USEC concluded that there were no credible accidents in which there wvuld be a simultaneous release of enough uranium and water to the scale pits to result in a criticality with a 6-inch Raschig ring depth. In responding to 8',7 :1 USEC considered and analyzed the common-mode failure in which a UF, release produces sufficient heat (i.e., the heat of reaction) to melt the fusible links on the fire sprinkler system heads. Such a condition could clearly make a criticahty in the scale pits likely.

USEC assumed the ERP room volume to be 25,000 ft). Based on his knowledge of ERP, the NRC Senior Resident inspector agreed with this assumption. Based on historical meteorological data, USEC conservatively assumed the temperature to be 95 degrees F and the relative humidity to be 60%. According to USEC, under these conditions, the water content in ERP would be the highest at a density of 0.7 g/ft'. The NRC staff independently validated 8 ,

i o .

this value by using the Psychrometric chari $ By assuming the maximum stoichiometric amount of gas phase UF. (175 kg) reacting instantaneously with the entire inventory of atmosphene moisture in ERP (17.5 kg), USEC determined the amount of heat generated to be i about 50,000 kJ.

Since UF. cannot exist in liquid form below its triple point of 22 psia, a rupture of a cylinder containing liquid UF. results in instantaneous expansion into both solid and gaseous fctms, the fraction of each depending on the process conditions (pressure, temperature, etc.) and the release process (release rate) of the liquid UF.. USEC conservatively assumed the liquid UF.

release of 386 pounds to form 62% vapor and 38% solid. This would result from UP. at a temperature of about 240 degrees F whereas product UF is withdrawn at a temperature of about 160 degrees F for which the vapor fraction would be reduced to about 45%". USEC accounted for the heat of sublimation of the solid fraction (147 lbs) of the released UF.. Based on a heat of reaction of 145 Btullb UF., USEC calculated the maximurn temperature nse to be from 95 degrees F to 194 degrees F. In addition, USEC assumed that gaseous UF.would preferentially react with moisture and that the solid UF. fraction would first sublime and then react with the moisture. The heat of reaction of solid UF. is significantly higher than that of vapor. However, based on the Gibbs Free Energies of all constituents, the staff confirmed that UF. vapor would preferentially react with the moisture. The Gibbs Free Energies for UF. (gas) and UF. (solid) reacting with water vapor are 138 kJ and 133 kJ, respectively. The staff also confirmed other data and assumptinns used by USEC and agraed that based on a 386 pound UF. release, the temperature would not rise more than 99 degrees F for the caso evaluated.

Accounting for heat transfer / loss to the structures and equipment contained in ERP and to the outside environment and assuming a lower vapor release fraction, would significantly reduce this temperature increase.

The staff notes that a Martin Marietta report provides a plot of the final temperature of a compartment into which UF. is released. Based on that plot, an instantaneous UF liquid release of 500 pounds at 148 degrees F yielding 42% vapor, into an air tight 25,000 ft' compartment at i atm and 80*F with a relative humidity of 60%, would raise the temperature by about 70"F. The graph shows that releases lower than 500 lbs sharply reduce the rise in temperature and that releases greater than 500 pounds tend to slightly reduce the rise in temperature indicating 500 pounds to be the worst case release from a temperature increase stand point at ERP. Another graph shows the pressure rise for 500 lb and 10,000 lb releases in the same compartment to be 2 psi and 5 psi, respectively. This indicates that following a large UF release in ERP, a significant convective outflow of the heated atmospheric components in ERP through the openings in the doors would occur.

Based on an independent assessment of USEC's assessment of heat generated in ERP during a UF release, the staff concludes that the likelihood of actuating the sprinkler system in ERP from a UF. release is very small.

Safeguards and Ee.Cutily The staff has not identified any safeguards or secunty related implications from the proposed amendment. ,

9 w_

~

, EIMRQHMENTAL REVIEW lssuance of the requested amendment to the Portsmouth Certificate of Compliance (GDP 2), to amend the Withdrawal Stations Standby Operational Mode definition, is subject to the categorical exclusion provided in 10 CFR 51.22(c)(19) and will not have a significant impact on the human environment. Therefore, in accordance with 10 CFR 51.22(b), neither an environmental assessment nor an environmental impact statement is required for the proposed action.

CONCLUSION Quantities of uranium could accumulate in the scale pits if a UF release were to occur in ERP, Should this accumulated mass of uranium become sufficiently moderated, a criticality is possible. However, based on an independent criticality safety review, the staff has determined that reducing the minimum depth of Raschig rings in the scale pits, as requsated by USEC, would not significantly raise the risk of criticality. The staff's safety determination was based on three primary areas of review: 1) Adequacy of contingency analysis for all credible process upsets; 2) Reliability of the controls; and 3) Adequacy of controlindependence (common mode failures). The basis for the staff's conclusion is based on the following controls and requirements:

a. To maintain the integrity of the UF,, pressure boundary, which provides geometry and mass control, USEC is committed to applying appropriate quality assurance requirements to process gas piping and equipment (including valves),

b. To provide moderation control, scale pits are inspected weekly for the presence of liquids. Any liquid found, is transferred out of the scale pits appropriately,

c. Maximum uranium enrichment of 10% is ensured by the use of in-line gamma and mass spectrometers or via samples if the spectrometers are not operational.

d. Raschig rings in the scale pits are inspected for settling and damage at least on an annual basit USEC is also committed to maintain the Raschig rings according to other requirements of ANSI /ANS 8.5 entitled "Use of Borosilicate-glass Raschig Rings as a Neutron Absorber in Solutions of Fissile Material"

e. The scale pits are required to be maintained free of uranium buildup.

f. To prevent recirculating cooling water (RCW) from entering the coolant system, the pressure of the RCW is maintained at least 5 psilower than the coolant system. A pressure switch is provided to automatically trip the UF. withdrawal compressor if this minimum pressure differential requirement is not maintained.

g. Smoke detectors are provided in ERP to monitor for UF releases. A UF out-leakage detection system has the capability of automatically isolating the pigtailif two smoke deiector heads detect smoke at the withdrawal station. While these lo

P

h. The maximum UF. pressure at the ERP station is maintained below 60 psia.

l. Prior to withdrawing UF into a product cylinder, a cold pressure check of the cylinder is performei The cylinder is rejected if the pressure is greater than 10 inches of mercury which provides indication of the presence of moderator or a hydrocarbon which can explosively react with UF.. The cylinder is also visually inspected for damage and weighed before being attached to the pigtail.

[ The staff independently reviewed and found acceptable, USEC's assumptions and calculations leading to the conclusion that for a large UF release in ERP, the heat generated by the exothermic reaction of UF with water vapor in ERP will not be sufficient to actuate the sprinkler system which could introduce moderator into the scale pits. .

k. There is a specific coolant pressure TSR Safety Limit (SL) of 440 psig. The purpose of this limit is to prevent the over pressurization and rupture of the coolant system which could result in the subsequent release of UF due to over pressurization and subsequent rupture of the UF containment boundary.

l. There are specific TSR Limiting Conditions of Operation (LCOs), Action Statements for conditions where LCOs are exceeded, and Surveillance Requirements (SRs), dealing with minimum number of operable smoke detectors / alarms to detect and indicate a release of UF.; coolant high pressure relief to ensure that the TSR SL of 440 psig is not exceeded; pigtailisolation system to limit the UF. release to less than 127 pounds in case of a pigtail failure; assay monitoring to ensure that the TSR specified maximum assays for the accumulators and cylinders are not exceeded; cylinder cart movement restrictions to ensure that a cylinder is not moved while it is connected to the withdrawal manifold; liquid UF. cylinder movement methods and restrictions to minimize the risk of a liquid UF. cylinder drop and rupture; UF. cylinder weight monitoring to ensure that the TSR specified fill weights are not exceeded; and restrictions on heating solidified UF. plugs to prevent pipe rupture that could be caused by local liquefaction and expansion.

m. There are specific general design feature requirements and associated SRs related to the design, cons action, testing and maintenance to ensure that the intended functions of UF, cylinders and pigtails are met so that they do not fail during normal operations; cylinder lifting cranes and fixtures to ensure that a cylinder is not dropped and ruptured; and Raschig rings in scale pits to enhance criticality safety.

Based on the primary reasons provided above, the staff finds it acceptable for USEC to reduce the depth of Raschig rings from 12 to 6 inches at ERP withdrawal stations 1 A and 2. The chosen control parameters of mass and moderator are found adequate, in addition, control independence has been demonstrated by USEC for a postulated common mode failure in which a UF release may produce sufficient heat to activate the fire sprinkler system. It is clear from the results reported that the neutron poison (in the form of borosilicate glass Raschig 11

, ring 3)is a tertiary control, and therefore, the reduction of nng depth does not affect the safety basis of this operation.

Based on the information provided in this CER, the NRC staff approves and grants this amendment. Region lli staff have no objection to this proposed action.

Attachment:

Figure 1 ERP UF. Flow Diagram hincipal Contnbuton Yawar Faraz Jack Davis Don Stout REEERENCES

1. J.H. Miller to C.J. Paoeriello, 'Portsmouth Gaseous Diffusion Plant (PORTS) Certificate Amendment Request Scale Pit Raschig Rings," GDP 97 0075, May 16,1997.

2. TID 7028, Cntical Dimensions of Systems Containing!35U. mPu. and mU, H.C. Paxton, et.

al, United States Atomic Energy Commission, Division of TechnicalInformation, June 1964.

3. TID 7016, NudgaLSafety Guide, Revision 2, J.T. Thomas, ORNL, June 1978.

4. ARH 600, Criticality Handbonk, R.D. Carter, et. al, Atlantic Richfield Hanford Company, June 30,1968.

5. L.M, Petrie and N F. Landers.

• KENO V a: Ar, improved Monte Carlo Coticality Program with Supergrouping," ORNL/NUREG/CSD 2, Revision 4, Volume 2, Section F11, ORNL, 1985.

6. NUREG/CR-0200, SCALE: A Modular Code _ System for Performing Standardized CDmputer Analyses for Licensina Evaluation, Revision 4, Apnl 1992.

7. R. Kinsey and B.A. Magumo," Data Formats and Procedures for the Evaluated Nuclear Data File," ENDF, BNL NCS 50496, November 1983.

8. PORTS Safety Analysis Report, Chapter 4, Accident Analy113, September 15,1995.

9, ORNL/TM 12292, Estimated Critical Conditions foLMQ:E2-H2O Sy11 ems in Fully Water-Reflected Spherical _ Geometry, W.C. Jordan, J.C. Tumer, ORNL, December 1992.

12

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

e s

10. POEF 520 95172, KENO V.a Analysis of the X_226 Extended Ranoe Product (ERP) i Withdrawal Station Scale Pit in Support of NCSA 0326-015.001, D.J. Lindenschmidt, Battelle Memorial institute, September 18,1996.

11. ORNUCSD/TM 284, Minimum Man _qf Moderator Reauired for Criticality of Homoneneous Low Enriched Uranium Systemt, W.C. Jordan, J.C. Tumer, ORNL, December 1992.

12. 1993 Ashree Handboo.8, Fundamentals, American Society of Heating, Refrigerating and Air Conditioning Engineers, Copyright 1993.

13. Perry's, Chemical Enaineers' Handbook, Sixth Edition, McGraw Hill Book Company, ,

1984.

14, Finis S. Patton, et. al, Enriched Uranium ProCJssina, The Macmillan Company,1963.

15. Gordon J. Van Wylen, et. al., Fundamentals of Classical Thermodynamics, Wiley, 1978.

16. M. Siman-Tov, et. al., Scenados and AnalvticeW9_thods for UF. Releases at NRC-(dpensed Fuel Cvele Facilities, NUREG/CR 3139, ORNUENG/TM 25,1984 1

1 DISTRIBUTION: (Control No.130S)

Docket 7o 7002 NRC File Center PUBLIC CCox, Rlli KO' Brian, Rlli NMSS t/f NMSS dir. ofc. r/f FCSS t/f SPB t/f PHiland, R:ll OStout, FCLB DHartland, Rlil KWinsberg, OGC WSchwink, FCOB JDavis, FCOB G:\RCHGCER.YHF OFC SPB [ hPB FCOB T MPB S4h NAME YFara C Hoadley ( JDavis tin R le\on DATE Y /2//97 $ /N /97 2 /7d/97 / /9 / 19 7 C = COVER E = COVER & ENCLOSURE N = NO COPY OFFICIAL RECORD COPY 13

E o .

a. j I

W I

< ....i. i

=

"""' i -

, 9 ac.ia fy/.m e s pes.

g)a

f. r r

y=.m .

r > c e.,,i, we

} ..

,si Pt.e. II #8'8 8.f ets.esas..

() HPV a, 9

+ 3* y!

O.....

4 4 *d i

.v...

.i.,.

Y lpunc........v...

Pe...Rftige.ii...==pE \

s o.i.. , ) v e, c .e....e  ;

Lit.is A ' "

4.....i....

,. (

G-' s..e s..., Lt..is bCgCDd p.... v ei ..

. i..o ,...i., v....

, [ ] . . . . , ,. . i . . . ,

y ..... 3 w usi.e or.rei.. v.i..

ii. Aie op e.i.4 c..ie.i v.i..

(/ .

& H..d.or ..ied v.t..

r,..... e ,,i.... A v.i.. o ,..

.:. v.i.. ci.i..

(Pipe diameters and pressures are approximate)

Figure 1 ERP and LAW UF6 Flow Diagram

_.